SDR Forum Technical Conference 2007 Proceeding of the SDR 07 Technical Conference and Product Exposition

“The authors represent that the work is original and they are the author or authors of the work, except for material quoted and referenced as text passages. Authors acknowledge that they are willing to transfer the copyright of the abstract and the completed paper to the SDR Forum for purposes of publication in the SDR Forum Conference Proceedings, on associated CD ROMS, on SDR Forum Web pages, and compilations and derivative works related to this conference, should the paper be accepted for the conference. Authors are permitted to reproduce their work, and to reuse material in whole or in part from their work; for derivative works, however, such authors may not grant third party requests for reprints or republishing.”

Government employees whose work is not subject to copyright should so certify. For work performed under a U.S. Government contract, the U.S. Government has royalty-free permission to reproduce the author's work for official U.S. Government purposes.

A SURVEY OF MULTIPLE ANTENNA TX-RX METHODS FOR WIBRO/WIMAX

Seokho Kang (The School of Electrical, Computer and Communication Engineering, Hanyang University, Seoul, Korea; [email protected]); Jaeho Chung (Network Infra Laboratory, KT, Seoul, Korea; [email protected]); Seungwon Choi (The School of Electrical, Computer and Communication Engineering, Hanyang University, Seoul, Korea; [email protected])

ABSTRACT which compared to single input single output (SISO). On the other hand, because SM scheme transmits different WiBro/WiMAX is an OFDMA-based multiple access signals from each transmit antenna, data-rate is increased service for wireless broadband multimedia environment. proportionally to number of transmit antennas. But SM Multiple antenna system has been considered to increase scheme does not provide a diversity gain. the capacity, service area, and data rate for WiBro/WiMAX In this paper, the performance of many MIMO base station system. In this paper, we compare the methods schemes and algorithms is compared and analyzed in for multiple antenna system of WiBro/WiMAX and show OFDMA-based wireless communication standard, the performance benefits of various multiple antenna WiBro/WiMAX downlink environment. Based on IEEE methods. We examine the space time transmit diversity 802.16-2004 [2] and 802.16e-2005 [3], WiBro (Wireless (STTD) and spatial multiplexing scheme for Broadband) is the next generation wireless communication WiBro/WiMAX. We first describe the signal modeling and standard that has been adopted in Korea. operations for STTD, spatial multiplexing. Then we This paper is organized as follows. In section 2, we explain advantages and disadvantages in each method, to examine the signal modeling of STTD, SM schemes and summarize the differences among the methods. We also algorithms. In section 3, we examine the structure of demonstrate various simulation results to show the WiBro/WiMAX downlink system. Simulation results to performances of STTD and spatial multiplexing which are apply many MIMO schemes are shown in section 4. Section compared in various communication circumstances. 5 concludes this paper.

2. MIMO SCHEMES 1. INTRODUCTION 2.1. Space Time Transmit Diversity Wireless communication using multiple input multiple output (MIMO) has recently emerged as one of the most 2.1.1. 2×1 Space Time Code significant technical issues to increase the system capacity There is MIMO system that has two transmit antennas and with limited resources in modern communications. MIMO a single receive antenna in figure 1. If two symbols to be system has many benefits that increase the system transmitted are s1 and s2 , STC sends the following for reliability, increase the data-rate and enlarge the wireless two symbol times: service coverage. However, all MIMO systems have not Antenna 1 2 these benefits. Each MIMO scheme has advantages and Time 0 disadvantages compared to one another. s1 s2 There are largely two algorithms of MIMO schemes: * * 1 - s2 s1 one is space time transmit diversity (STTD) and the other is During two symbol times, transmitted symbols in each spatial multiplexing (SM). STTD scheme transmits the antenna are orthogonal with each other. Using matrix orthogonal signals in each transmit antenna and is often expression, received symbol vector is expressed in called by Alamouti code or space time code [1]. This following (1). An assumption is that the channel h , h scheme can obtain a diversity gain like the maximum ratio 1 2 remain constant over two symbol times. combining in receive antenna, but data-rate is not increased

* - s2 s1 s s 2 Data 2 1 h ˆ ˆ STC y2 y1 STC s2 s1 Data Modulation Encoder ant 2 Decoder Demodulation

* s1 s2 space

time

Fig. 1. 2×1 space time code

ant 1 h1 * - s s1 2 h2 y3 y1 STC s s 2 1 STC Decoder ant 2 sˆ2 sˆ1 Encoder h3 Combining * y4 y2 s1 s2 h4 STC Decoder space

time

Fig. 2. 2×2 space time code

* 2 2 é y1 ù é s s ùéh1 ù én1 ù és ù éh h ùéy ù é(h + h ) × s + N ù = 1 2 + (1) 1 3 4 3 3 4 1 3 (4) ê ú ê * * úê ú ê ú ê ú = ê * úê *ú = ê 2 2 ú y - s s h n s h - h y ë 2 û ë 2 1 ûë 2 û ë 2 û ë 2 û ë 4 3ûë 4 û ëê(h3 + h4 )× s2 + N4 ûú where n , n are additive white Gaussian noise. ~ 2 2 2 2 1 2 sˆ s s é ù é 1 ù é 1 + 1 ù ( h1 + h2 + h3 + h 4 ) × s1 + N1 + N3 (5) Received symbols are decoded as below. ê ú = ê~ ú = ê 2 2 2 2 ú sˆ s + s ( h + h + h + h ) × s + N + N 2 2 ë 2 û ë 2 2 û ëê 1 2 3 4 2 2 4 ûú ésˆ ù éh* h ùé y ù é( h + h )×s + N ù 1 1 2 1 1 2 1 1 (2) We can see that data-rate of 2×2 STC is the same with ê ú = ê * úê * ú = ê 2 2 ú sˆ h - h y ë 2 û ë 2 1 ûë 2 û ëê( h1 + h2 )×s2 + N2 ûú 2×1 STC. As (2), we can see that 2×1 STC obtains a diversity gain like the maximum ratio combining (MRC) of 1×2 2.2. Spatial Multiplexing receive diversity system. In summary, the 2×1 STC achieves the same diversity order and data-rate as a 1×2 SM is a scheme that a set of independent symbols are receive diversity system with MRC but with 3dB penalty, transmitted in each transmit antenna, which make the data- owing to the redundant transmission required to remove the rate be doubled. Mathematical expression is as follows. spatial interference at the receiver. y = Hx + n , i.e. é y ù é h h L h ù é x ù é n ù 2.1.2. 2×2 Space Time Code 1 11 12 1N 1 1 (6) ê y ú ê h h L h ú ê x ú ê n ú Figure 2 is 2×2 STC and STC encoding is operated as ê 2 ú = ê 21 22 2N ú ê 2 ú + ê 2 ú follow. ê M ú ê M M O M ú ê M ú ê M ú ê ú ê ú ê ú ê ú é y y ù é s s ùéh h ù én n ù ëyM û ëhM1 hM 2 L hMN û ëxN û ënM û 1 3 = 1 2 1 3 + 1 3 (3) ê ú ê * *úê ú ê ú where M is the number of receive antennas, N is the ëy2 y4 û ë- s2 s1 ûëh2 h4 û ën2 n4û number of transmit antennas and 2 . · term 2×2 STC decoding is operated like (2) in each receive n ~ Nc (0,s I) antenna and then, decoded results are combined as below. indicates a matrix, · term indicates a vector. ~s é * ù y é 2 2 ù é 1 ù h1 h2 é 1 ù ( h1 + h2 ) × s1 + N1 , Figure 3 is the structure of 2×2 SM. In SM scheme, to ê ú = ê * úê * ú = ê 2 2 ú ~s h - h y detect received symbols, the detection processing is ë 2 û ë 2 1ûë 2 û ëê( h1 + h2 ) × s2 + N2 ûú required.

h x1 21 y1 2 h1 yˆ yˆ Data x2 x1 MIMO 2 1 Data S / P ant 2 ant 2 Modulation h22 Detection Demodulation x 2 y2

Fig. 3. 2×2 spatial multiplexing

There are many algorithms that are used to detection ~ 2 H H H -1 W = arg min Eé Wy - x ù = H (HH + E[nn ]) processing and in this paper we examine maximum W ëê ûú (9)

likelihood (ML), zero forcing (ZF), minimum mean square ~ H H H -1 error (MMSE), and successive interference cancellation xˆ = W ×y = H (HH + E[nn ]) y (SIC), ordered SIC (OSIC). Because noise effect is considered, its performance is better than ZF. But the computational complexity is larger 2.2.1. Maximum Likelihood than ZF. The following equation is criterion for ML detection. 2 2.2.4. Successive Interference Cancellation xˆ = argmin y - Hx (7) x The earliest known SM receiver was invented and where x is a transmitted symbol vector, y is a received prototyped in Bell Labs and called Bell Lab layered space/time (BLAST) [4]. Like other SM MIMO systems, symbol vector, H is a channel matrix and xˆ is a estimated BLAST consists of parallel “layers” supporting multiple transmit symbol vector. simultaneous data streams. The layers (substreams) in If channel information is known at the receiver, BLAST are separated by interference cancellation estimated transmit symbol vector is obtained when x is techniques that decouple the overlapping data streams. The most important technique is vertical BLAST (V-BLAST) substitute for all possible input vector. But if the order of and we examine subsequent interference cancellation, SIC. the modulation (e.g., 4 for QPSK) is large, the SIC detection process is as follow. computational complexity is very high. Therefore ML 1: H = [h h L h ] detection is very optimal but has a disadvantage of 1 2 N T computational complexity. + H -1 H 1T 2T N T 2 : H = (H H) H = [w1 w1 L w1 ] 1 ~ 1T 2.2.2. Zero Forcing 3: x1 = w1 y As in other situations in which the optimum decoder is an ~ 4 : xˆ1 = Q(x1 ) (Quantization) intolerably complex ML detector, a sensible next step is to 5 : y = y - xˆ h (Interference cancellation) consider linear detectors that are capable of recovering the 2 1 1 6 : H = [h h L h ] transmitted vector x . The most obvious such detector is 2 2 3 N -1 T T T T the zero forcing detector, which sets the receiver equal to 7 : H + = H H H H H = w 2 w3 L w N 2 ( 2 2 ) 2 [ 2 2 2 ] the inverse of the channel, or more generally to the pseudo- T 8: ~x = w 2 y inverse. 2 2 2 ZF detection is as below. ~ 9: xˆ2 = Q(x2 ) + + H -1 H xˆ = H y with H = (H H) H (8) 10: y = y - xˆ h 3 2 2 2 where upper subscript H is complex transpose operation. M ZF detection is easily operated because of the pseudo- In each layer, the transmitted symbol is estimated inverse of channel matrix, but the bad spatial subchannles based on zero forcing algorithm and in next layer can severely amplify the noise and its performance is not interference is removed from received symbol vector. SIC good. algorithm is progressed in the order of layers.

2.2.3. Minimum Mean Square Error 2.2.5. Ordered Successive Interference Cancellation A logical alternative to the zero forcing detection is the Unlike SIC, the ordered SIC is performed in the order of MMSE detection and the algorithm is as follow. decreasing SNR. This algorithm is as follow.

Cluster Randomizat Bit Data Data & Pilot Subcarrier IFFT Add FEC Repetition Renumberi Tx Ant#1 ion Interleaver Modulation Allocation Permutation (1024p) CP ng

MIMO Encoding

Cluster Pilot Pilot Data & Pilot Subcarrier IFFT Add Renumberi Tx Ant#2 Generation Modulation Allocation Permutation (1024p) CP ng

Subcarrier Cluster FFT Remove De- De- Rx Ant#1 (1024p) CP permutation renumbering

De- Bit Data Channel MIMO Channel randomizati De- Summing De- Decoding Decoding Estimation on interleaver modulation

Subcarrier Cluster FFT Remove De- De- Rx Ant#2 (1024p) CP permutation renumbering

Fig. 4. WiBro/WiMAX MIMO downlink block diagram

y = Hx + n H = [h1 h 2 L h N ] Table 1. Simulation parameters T initialization : G = H + G = [g1 g 2 L g N ] Parameters Value 1 i i i i Subcarrier Permutation Mode PUSC i =1 Carrier Frequency 2.3 GHz y = y Channel Bandwidth 8.75 MHz 1 2 FFT Size 1024 point recursion : k = arg min g j i jÏ{k1 ,...,ki-1} i Number of Data Subcarriers 720 Number of Pilot Subcarriers 120 w = g ki ki i TDD Frame Length 5ms ~ T CP(Cyclic Prefix) Ratio 1/8 x = w k y ki i i Doppler Frequency 127 Hz xˆ = Q(~x ) (Quantizat ion) ki ki Data Modulation QPSK Channel Coding Convolutional Coding y = y - xˆk h k (Interference cancellation) i+1 i i i (Coding Rate) (1/2) + G = H (H = H with columns k1,...,ki , ki set to 0) i+1 ki ki Number of Tx Antennas 2 i = i +1 Number of Rx Antennas 1 or 2 Number of Used Slots 20

3. STRUCTURE OF WIBRO/WIMAX SYSTEM Under WiBro/WiMAX MIMO downlink environment, Based on IEEE 802.16-2004 and IEEE 802.16e-2005, computer simulation is progressed on many MIMO WiBro/WiMAX standard specification has options about schemes. There are simulation parameters on Table. 1. An MIMO schemes. When MIMO schemes are applied to assumption is that channel information is perfectly known WiBro/WiMAX system, the system has advantages of at receive antenna. increasing system capacity and spectrum efficiency. Figure In case of 2×1 and 2×2 STTD, figure 5 is simulation 4 is block diagram of WiBro/WiMAX MIMO downlink result. As shown in figure 5, if the system has more system. transmit antennas and receive antennas, it obtains a In the specification of WiBro/WiMAX, STTD and SM diversity gain like MRC of diversity system. are optionally supported. For example, in the circumstance Figure 6 is a simulation result of 2×2 SM for many SM having good channel quality, SM may be selected and in schemes and figure 7 is a throughput curve of ML, ZF, cell edge circumstance having poor channel quality, STTD MMSE, SIC, OSIC. As shown in figure 6, ML detection is may be selected. the most optimal and based on zero forcing, SIC and OSIC seem to be better performance than ZF detection. But 4. SIMULATION RESULTS MMSE has a better performance than SIC and OSIC. In figure 7, when applying many SM schemes, their

0.1 300 )

s 250 p b k ( t R u

E 0.01 p 200 B h g u o r h

T 150 1x1 0.001 ML 100 ZF MMSE 50 SIC OSIC 0.0001 0 0 5 10 15 20 25 0 3 6 9 12 15 18 21 24 27 30 SNR SNR Fig. 5. WiBro/WiMAX STTD downlink performance Fig. 7. WiBro/WiMAX 2×2 SM downlink data throughput

1 ML ZF In this paper, we have investigated the MIMO MMSE SIC algorithms which are interested for commercial application. 0.1 OSIC MIMO technology is accepted as the standard for multiple antenna scheme of WiBro/WiMAX system. This paper analyzed the performance of MIMO under WiBro/WiMAX

R signal environments through computer simulations. E 0.01 B 6. ACKNOWLEDGEMENT

0.001 This work was supported by the HY-SDR Research Center, Hanyang University, Seoul, Korea, under the ITRC program of MIC, Korea and was partly supported by the IT R&D program of MIC/IITA [2007-S001-01, Advanced 0.0001 6 9 12 15 18 21 24 MIMO system]. SNR Fig. 6. WiBro/WiMAX 2×2 SM downlink performance 7. REFERENCES

[1] S.M. Alamouti, “A simple transmit diversity technique for throughput is doubled than SISO system. This result wireless communications,” IEEE Journal on Selected Areas coincidences with mentioned before in this paper. in Communications, Vol. 16, No. 8, pp. 1451-1458, October 1998. [2] IEEE 802.16-2004, “IEEE Standard for Local and 5. CONCLUSION metropolitan area networks, Air Interface for Fixed Broadband Wireless Access Systems,” October 2004. In this paper, STTD and SM schemes leading MIMO [3] IEEE 802.16e-2005, “IEEE Standard for Local and technology are viewed from signal modeling and metropolitan area networks, Air Interface for Fixed and Mobile Broadband Wireless Access Systems, Amendment simulation performance of WiBro/WiMAX wireless 2:Physical and Medium Access Control Layers for Combined communication environment. As shown in simulation Fixed and Mobile Operation in Licensed Bands,” February results, in a good channel circumstance, we prefer the SM 2006. scheme due to increase data-rate. On the other hand, in an [4] G.J. Foschini, “Layered space-time architecture for wireless adverse channel circumstance, the STTD scheme seems to communication in a fading environment when using multiple antennas,” Bell Labs Technical Journal, Vol. 1, No. 2, pp. be better due to obtain a diversity gain. 41-59, 1996.