Chapter 4 SC-OFDM-OFDMA SDR Architecture
4.4 Computer Simulations
In this section, computer simulations are conducted to evaluate the performance of the proposed SC-OFDM-OFDMA SDR system. Also, the computer simulations of the proposed SDR system with transmit techniques such as STC and AAS are shown in this section. Throughout the simulations, we only deal with discrete time signal processing in the baseband, hence pulse shaping filter is removed from consideration for simplicity.
Figures 4.6, 4.7 and 4.8 show the proposed SDR receiver architectures for OFDM, OFDMA, and SC transmission mode. As seen in these three figures, it is clearly observed that most functional blocks of the three modes are used commonly. So the proposed SDR receiver can be switched among the three modes via the concept of SDR operation. Figures 4.6, 4.7 and 4.8 also give the parameter adjustments among the three modes.
Figure 4.6: SDR receiver architecture for OFDM mode
Time/Frequency Joint design of
timing &
extract needed subchannels
& pilot tones Subchannels &
Pilot tones extraction
Window size: v=CP length/2 f=FFT length
l = even numbered symbols l = odd numbered symbols Residual freq.
estimation
Adjustable parameter
extract needed subchannels
& pilot tones
l = even numbered symbols l = odd numbered symbols Residual freq.
Figure 4.7: SDR receiver architecture for OFDMA mode
Time/Frequency Joint design of
timing &
Window size: v=CP length/2 f=FFT length
Figure 4.8: SDR receiver architecture for SC mode
In the simulation results, BER performance as a function of Eb/N0 is evaluated. useful part of the received data,
Nc TFFT
Tsamp is sampling time, Nt is the number of transmit
antennas, M equals B is the bandwidth of the
system, is the number of data subcarriers, and E
( )
log2 constellation size of modulation ,
c data_
N b is the bit energy. If
equals , we can obtain Equation (4.11). Total transmit power is normalized to 1.
Hence, we can obtain the relation between noise power and E
c data_
With the consideration of practical implementation, we evaluate the performance impact due to the proposed jointly designed synchronization algorithm and channel estimation algorithm under SUI-1 channel and Vehicular A channel. The parameters of multipath fading channel models for SUI-1 channel and Vehicular A channel are shown in Table 4.8. SUI-1 channel model is mainly used in SC mode. Under Vehicular A channel model, the individual multi-path is subject to the independent Rayleigh fading, whose time domain correlation is implemented by the Jakes model. We also consider the vehicular speeds at 3 km/h and 120 km/h in Vehicular A channel model. All BER are evaluated by averaging over 30000 frames, and each frame has 10 symbols.
Table 4.8: Channel models used in the simulations
SUI-SUI-1 channel1 channel K=20,10,0
SUI-SUI-1 channel1 channel K=20,10,0
Table 4.9 lists all parameters used in our simulations. The symbol time and the subcarrier spacing are fixed. The oscillator offset is set to be 5 KHz for carrier frequency at 2.5 GHz. The CP length is chosen as to ensure the maximum delay spread smaller than the CP length, and the modulation schemes used here are BPSK, QPSK, 16QAM and 64QAM. The size of FFT is scaled according to the different bandwidths to keep the subcarrier spacing constant. Table 4.10 gives all the scalability parameters used for the different FFT sizes and bandwidths.
b/ 4 T
Table 4.9: Simulation parameters for mobile WiMAX
1.25/2.5/5/10/20 MHz Channel BW
Channel BW
10 8
Data Symbols Per 10 Data Symbols Per
Frame
Parameters Mobile WiMAXMobile WiMAX
SUI-1 (for SC)/
ITU VehA 3, 60, 120 km/hr (for other
modes)
ITU VehA 3,60,120 km/hr
2/1 STCSTC
ITU VehA 3,60,120 km/hr
Data Symbols Per 10 Data Symbols Per
Frame
Parameters Mobile WiMAXMobile WiMAX
SUI-1 (for SC)/
ITU VehA 3, 60, 120 km/hr (for other
modes)
ITU VehA 3,60,120 km/hr
2/1 STCSTC
ITU VehA 3,60,120 km/hr
4/1 AASAAS
Table 4.10: OFDMA scalability parameters for different bandwidth
BER performances of the proposed SC-OFDM-OFDMA SDR architecture in OFDM transmission mode are shown in Figures 4.9, 4.10, and 4.11. In these simulations, the size of FFT equals 256 and the system is equipped with a single antenna. In OFDM transmission mode, the short preambles can be used to perform timing and frequency synchronization. It is obviously observed that the BER performance with preamble-aided channel estimation is better than that with pilot-aided channel estimation at 3 km/h under Vehicular A channel model because its number of pilots is not enough to track the channel variation between the two adjacent pilot subcarriers. However, under Vehicular A 120 km/h channel, the performance with pilot-aided channel estimation is superior to that with preamble-aided channel estimation. With pilot-aided channel estimation, the BER curve will be flat at high SNR due to the channel estimation error.
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
256-FFT OFDM with coded BPSK
All ideal
ITU VehA 3km/h (preamble) ITU VehA 3km/h (pilot) ITU VehA 120km/h
Figure 4.9: BER performance with 256-point FFT with BPSK in OFDM transmission mode under Veh A channel
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/N
o
BER
256-FFT OFDM with coded QPSK
All ideal
ITU VehA 3km/h (preamble) ITU VehA 3km/h (pilot) ITU VehA 120km/h
Figure 4.10: BER performance with 256-point FFT with QSPK in OFDM transmission mode under Veh A channel
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/N
o
BER
256-FFT OFDM with coded 16QAM
All ideal
ITU VehA 3km/h (preamble) ITU VehA 3km/h (pilot) ITU VehA 120km/h
Figure 4.11: BER performance with 256-point FFT with 16QAM in OFDM transmission mode under Veh A channel
BER performances of the proposed SC-OFDM-OFDMA SDR architecture in OFDMA transmission mode are shown in Figures 4.12-4.15. The FFT size in OFDMA transmission mode can be scaled to be 128, 512, 1024, and 2048. In OFDMA transmission mode, the CP can be used to perform timing and frequency offset synchronization. It is worthy of mention that the proposed SDR system keeps the symbol time and subcarrier spacing constant. With these mentioned practical considerations, the overall BER performance of the OFDMA-2048 mode with QPSK will have about 0.3 dB implementation loss compared with the ideal case under Vehicular A 120 km/h channel model. Therefore, with the proposed joint design of synchronization algorithm and the channel estimation algorithm, the proposed SDR system performance has very little degradation compared with the ideal case system.
First, the performances of the different FFT sizes are compared. It is observed that from the OFDMA-2048 mode to the OFDMA-128 mode, the performance loss compared
with the ideal case becomes larger and larger (from 0.3 dB to 5 dB as BER=0.001). It’s found that the OFDMA-128 mode performs poorest than other FFT sizes because its CP length is too short to estimate the frequency offset accurately. Figure 4.16 figures out the MSE of the frequency offset estimate with 128-point, 512-point, 1024-point and 2048-point FFT under the same channel condition. Second, as seen in Figure 4.12, it is obviously observed that the BER curve will be flat due to the timing and frequency offset estimation error. Third, the sample time varies with the different bandwidths.
When the FFT size becomes larger, the sample time will become smaller and then make the delay spread larger. In the case with 128-point, 512-point, 1024-point and 2048-point FFT, it can be found that the delay spread which is smaller than the CP length will not degrade the performance although the delay spread becomes larger.
Finally, in OFDMA transmission mode, its number of pilot tones is enough so that the two adjacent pilots enable to track the channel variation accurately between the two adjacent pilot subcarriers.
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
128-FFT OFDMA, VehA 120 km/h
All ideal-QPSK QPSK 16QAM 64QAM
Figure 4.12: BER performance with 128-point FFT in OFDMA transmission mode under Veh A channel
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
512-FFT OFDMA, VehA 120 km/h
All ideal-QPSK QPSK 16QAM 64QAM
Figure 4.13: BER performance with 512-point FFT in OFDMA transmission mode under Veh A channel
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/N
o
BER
1024-FFT OFDMA, VehA 120 km/h
All ideal-QPSK QPSK 16QAM 64QAM
Figure 4.14: BER performance with 1024-point FFT in OFDMA transmission mode under Veh A channel
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
2048-FFT OFDMA, VehA 120 km/h
All ideal-QPSK QPSK 16QAM 64QAM
Figure 4.15: BER performance with 2048-point FFT in OFDMA transmission mode under Veh A channel
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/N
o (dB)
MSE (Hz)
MSE of frequency offset estimation with CP method
128-FFT 512-FFT 1024-FFT 2048-FFT
Figure 4.16: MSE of frequency offset estimates in OFDMA-128, 512, 1024 and 2048 mode
BER performances of the proposed SC-OFDM-OFDMA SDR architecture in SC transmission mode are shown in Figures 4.17 and 4.18. SUI-1 channel is used in these simulations and its parameter is shown in Table 4.8. The outer code is RS(240,208,16) for QPSK and RS(496,432,32) for 16QAM and the inner code is CC code rate 1/2.
Timing and frequency synchronization schemes are performed by utilizing its CP like the case in OFDMA transmission mode. It is obviously observed that LOS channels with Rice factor K=20 and K=10 are suitable for SC transmission mode. The performance of the case with Rice factor K=0 is poorer because the SC transmission mode is not robust against NLOS environments.
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
SC,QPSK,FDE-256
SUI-1,K=0 SUI-1,K=10 SUI-1,K=20
Figure 4.17: BER performance with coded QPSK in SC transmission mode under SUI-1 channel
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
SC,16QAM,FDE-256
SUI-1,K=0 SUI-1,K=10 SUI-1,K=20
Figure 4.18: BER performance with coded 16QAM in SC transmission mode under SUI-1 channel
Finally, the performances with transmit techniques such as STC and AAS will be verified to exhibit the improved performances. Simulation results with STC are shown in Figures 4.19, 4.20 and 4.21. Under Vehicular A 3 km/h channel environments, STC with two transmit antennas and one receive antenna provides over 5 dB gain than in SISO mode at a bit error rate of 0.0001. With higher vehicular speed, the BER curve will be flat because the time-varying fading rate is too fast. Under Vehicular A 60 km/h channel environments, it is clearly observed that the lower order modulation is more robust against time-varying channel environments.
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
OFDMA-2048,QPSK,STC(2Tx1Rx)
VehA 3km/hr VehA 60km/hr VehA 120km/hr SISO VehA 3km/hr
Figure 4.19: BER performance with STC (2Tx1Rx) and QPSK in OFDMA-2048 mode under Veh A channel
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/N
o
BER
OFDMA-2048,16QAM,STC(2Tx1Rx)
VehA 3km/hr VehA 60km/hr VehA 120km/hr SISO VehA 3km/hr
Figure 4.20: BER performance with STC (2Tx1Rx) and 16QAM in OFDMA-2048 mode under Veh A channel
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
OFDMA-2048,64QAM,STC(2Tx1Rx)
VehA 3km/hr VehA 60km/hr VehA 120km/hr SISO VehA 3km/hr
Figure 4.21: BER performance with STC (2Tx1Rx) and 64QAM in OFDMA-2048 mode under Veh A channel
Next, the performances of using AAS techniques are shown in Figures 4.22-4.27.
Channel estimation is performed by using the pilot aided scheme. AAS techniques can generally be classified as either switched beamforming or adaptive beamforming.
Because adaptive beamforming is generally more digital-processing intensive than switched beamforming, they tend to be more costly and complex. So switched beamforming is considered and simulated in this system. The transmitter is equipped with four antenna arrays and the receiver is equipped with one antenna. As seen in Figures 4.22, 4.24 and 4.26, AAS with 16 beams and 8 beams provide about 6 dB and 4 dB in performance gains under Vehicular A 3 km/hr channel environments. As the number of beams decreases, the radiation angle of each beam increases and the intended user may not be in the center of the main beam. As shown in Figures 4.23, 4.25 and 4.27, under Vehicular A 120 km/hr channel environments, it is obviously observed that the performance is poorer because the time-varying fading rate is so fast
that the weights computed over the preamble period are not suitable for the remaining symbols.
0 2 4 6 8 10 12 14 16 18 20
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
OFDMA-2048,QPSK,AAS(4Tx1Rx) under VehA 3km/hr
16beams VehA 3km/hr 8beams VehA 3km/hr 4beams VehA 3km/hr SISO VehA 3km/hr
Figure 4.22: BER performance with AAS (4Tx1Rx) and QPSK in OFDMA-2048 mode under VehA 3 km/hr channel
0 2 4 6 8 10 12 14 16 18 20
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
OFDMA-2048,QPSK,AAS(4Tx1Rx) under VehA 120km/hr
16beams VehA 120km/hr 8beams VehA 120km/hr 4beams VehA 120km/hr SISO VehA 120km/hr
Figure 4.23: BER performance with AAS (4Tx1Rx) and QPSK in OFDMA-2048 mode under VehA 120 km/hr channel
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
OFDMA-2048,16QAM,AAS(4Tx1Rx) under VehA 3km/hr
16beams VehA 3km/hr 8beams VehA 3km/hr 4beams VehA 3km/hr SISO VehA 3km/hr
Figure 4.24: BER performance with AAS (4Tx1Rx) and 16QAM in OFDMA-2048 mode under VehA 3km/hr channel
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/N
o
BER
OFDMA-2048,16QAM,AAS(4Tx1Rx) under VehA 120km/hr
16beams VehA 120km/hr 8beams VehA 120km/hr 4beams VehA 120km/hr SISO VehA 120km/hr
Figure 4.25: BER performance with AAS (4Tx1Rx) and 16QAM in OFDMA-2048 mode under VehA 120km/hr channel
0 5 10 15 20 25 30 10-7
10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/No
BER
OFDMA-2048,64QAM,AAS(4Tx1Rx) under VehA 3km/hr
16beams VehA 3km/hr 8beams VehA 3km/hr 4beams VehA 3km/hr SISO VehA 3km/hr
Figure 4.26: BER performance with AAS (4Tx1Rx) and 64QAM in OFDMA-2048 mode under VehA 3km/hr channel
0 5 10 15 20 25 30
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Eb/N
o
BER
OFDMA-2048,64QAM,AAS(4Tx1Rx) under VehA 120km/hr
16beams VehA 120km/hr 8beams VehA 120km/hr 4beams VehA 120km/hr SISO VehA 120km/hr
Figure 4.27: BER performance with AAS (4Tx1Rx) and 64QAM in OFDMA-2048 mode under Ve A 120km/hr channe
4.5 Summary
In this chapter, the proposed SC-OFDM-OFDMA SDR transmitter architecture is introduced first. We also introduce the procedures of determining the segmentation block size to fit the input data size of the encoder. At the receiver, the proposed SC-OFDM-OFDMA SDR receiver architecture via the concept of SDR is introduced.
All the functional blocks have to be modified to adapt to the different air interfaces.
Synchronization is first presented, which consists of the jointly designed timing and frequency synchronization algorithm. The jointly designed timing and frequency synchronization scheme proposed in Chapter 3 also needs to be modified to adapt to the different preamble structures. Then, channel estimation, phase estimation and residual frequency offset estimation are described in the rest of this chapter. After that, we highlight the multiple-antenna transmit techniques such as STC and AAS, implemented on the SDR system to gain the improved performances. Finally, we evaluate the performances of the joint design of SC-OFDM-OFDMA SDR system and confirm that it works reliably among the three modes.
Chapter 5 Conclusion
In this thesis, a jointly designed transceiver is developed for Single Carrier, OFDM and OFDMA under the SDR architecture. The jointly designed SC-OFDM-OFDMA SDR architecture is developed to support the various air-interface standards specified by IEEE 802.16-2005 on a single SDR platform. In this way, the system possesses as many common components as possible for these three modes, and the transmitter and receiver can be switched among the three modes via the operation of SDR concept.
In Chapter 2, the transmitter architecture and specification of IEEE 802.16-2005 WiMAX system have been introduced. In the rest of this chapter, we also introduce the multiple-antenna transmit techniques such as STC and AAS adopted in this system. In Chapter 3, we introduce two channel models: the first is the SUI channel model to form the fixed wireless channel environment and the second is the ITU channel model to form the mobile channel environment. After that, synchronization, channel estimation and phase estimation algorithms for this SDR system are established. In particular, we propose a jointly designed timing and frequency synchronization algorithm for the SDR architecture. The jointly designed timing and frequency synchronization scheme proposed can lower the computational complexity and obtain the reliable timing and
frequency offset estimates. We also use the phase estimators to perform residual frequency offset estimation. By this way, the residual frequency offset estimation scheme can co-work with the proposed jointly designed scheme to obtain a wider range and better accuracy of estimates.
In Chapter 4, we first review the concept of SDR. SDR technology facilitates the implementation of some of the functional modules by software in the MAC and PHY layers. This helps in building the reconfigurable software radio systems where dynamic selection of parameters for some software-defined functional modules is possible. Next, the transmitter and receiver architecture of the SC-OFDM-OFDMA SDR system are proposed. At the transmitter side, the data block size is determined by the FEC block, modulator, and the number of subchannels allocated. We describe three mathematical equations for the three modes to facilitate determining the uncoded data block size. At the receiver side, we focus on the three blocks such as timing and frequency synchronization block, channel estimation block, and phase estimation block. Although these algorithms have been mentioned in Chapter 3, they need to be modified to match the functions of the prescribed air interfaces by setting suitable parameters. In particular, the jointly designed timing and frequency synchronization scheme proposed is modified in accordance with different preamble structures: with short preamble and without short preamble. The simulations indicate the proposed timing and frequency synchronization scheme with the CP can work reliably and the performances are close to the ideal case as long as the CP length is long enough. After that, we generalize some results from the three modes and give the adjustable parameters for the three modes on the same SDR platform.
Finally, we study the multiple-antenna transmit techniques such as STC and AAS specified by IEEE 802.16-2005. Compared with the SISO transmission mode, STC transmit technique can improve the performance of the system significantly when the
SS is at low mobility. AAS can also obtain the improved performance through the use of more than one antenna elements at the BS. Both of the two transmit techniques have the same advantage of remaining only one antenna at the SS side. This key advantage can be utilized by increasing more cost and complexity only at the BS side without raising the complexity of user’s devices.
Bibliography
[1] IEEE Std 802.16e-2005 and IEEE Std 802.16-2004/Cor 1-2005, “Part 16: air interface for fixed and mobile broadband wireless access systems,” Feb. 2006.
[2] IEEE Std 802.16-2004, “Part 16: air interface for fixed broadband wireless access systems,” Oct. 2004.
[3] “Mobile WiMAX - part 1: a technical overview and performance evaluation,”
WiMAX Forum, Feb. 2006.
[4] “Mobile WiMAX - part 2: a comparative analysis,” WiMAX Forum, April 2006.
[5] John Hoadley and Al Javed, “Overview: technology innovation for wireless broadband access,” Nortel Technical Journal, No. 2, July 2005.
[6] A. Salvekar, S. Sandhu, Q. Li, M. Vuong and X. Qian, “Multiple-antenna technology in WiMAX systems,” Intel Technology Journal, Vol. 8, Aug.2004.
[7] S. Alamouti, “A simple transmit diversity technique for wireless communications,” IEEE Journal on Select Areas in Communications, Vol. 16, No.
8, pp. 1451-1458, Oct. 1998.
[8] Jin M. Ku, Pil K. Kim, Se J. Lee, Simon Shin, and Chung G. Kang, “On the performance of broadband mobile internet access system,” ISWPC2006, Jan.
2006.
[9] Shuenn Gi Lee, “Performance of concatenated FEC under fading channel in Wireless-MAN OFDM system,” Proceedings of the 19th International Conference on Advanced Information Networking and Applications, July 2005.
[10] Pangan Ting, Yung-Yih Jian, Chia-Tsung Lin, Yu-Tao Hsieh, Chi-Fang Li, Tsan-Huang Chen, Shuenn Gi Lee, Wei Ping Chuang, Chao-Kai Wen and Racy J-H Cheng, “Downlink BER simulation for IEEE 802.16e OFDM-PHY,”
Proceedings of 2005 International Symposium on Intelligent Signal Processing and Communication Systems, June 2005.
[11] IEEE 802.16a-03/01, “Channel models for fixed wireless applications,” June 2003.
[12] Recommendation ITU-R M.1225, “Guidelines for evaluation of radio transmission technologies for IMT-2000,” 1997.
[13] P. Y Tsai, H. Y. Kang, and T. D. Chiueh, “Joint weighted least squares estimation of carrier-frequency offset and timing offset for OFDM systems over multipath fading channels,” IEEE Trans. Vehicular Technology, Vol. 54, No. 1, Jan. 2005.
[13] P. Y Tsai, H. Y. Kang, and T. D. Chiueh, “Joint weighted least squares estimation of carrier-frequency offset and timing offset for OFDM systems over multipath fading channels,” IEEE Trans. Vehicular Technology, Vol. 54, No. 1, Jan. 2005.