High Data Rate 5Ghz WLAN Air Interface Based on Multiple Transmit Multiple Receive Antennas

Wireless World Research Forum (WWRF)

High Data Rate 5GHz WLAN air interface based on Multiple Transmit Multiple Receive Antennas

Karine Gosse, Stéphanie Rouquette, Alexandre Ribeiro Dias, Sébastien Simoens Motorola Labs Paris, Espace Technologique de Saint-Aubin 91193 Gif-sur-Yvette Cedex E-mail: [email protected]

Subject Area: WG 4: Smart antennas for NG systems

.1 Objectives of the required research

WLANs are currently seen as a necessary complement of the cellular infrastructure, in order to provide sufficiently high bit-rates in dense urban environments so as to cope with data/multimedia traffic needs. Current 5GHz WLAN air interfaces rely on the OFDM (Orthogonal Frequency Division Multiplexing) modulation, and provide a bit-rate up to 54Mbps. So as to prepare for future wireless broadband systems “beyond 3G” , and in order to optimize the usage of the 5GHz spectrum, new air interfaces for WLANs based on OFDM need to be investigated, with increased peak data rate and improved cell throughput. There are several ways to envisage this: in the short term, higher data rates can be obtained with 20MHz channel bundling so as to increase the system bandwidth, but this has the drawback to reduce the number of available channels. A longer term approach is to investigate the usage of higher frequencies, such as the 60GHz band, where large amounts of spectrum are available, but on the other hand, the severe propagation conditions require to revisit the usage of the obtained system. The strategy followed here is the optimization of the system spectral efficiency with Multiple Transmit and Multiple Receive (MTMR) antennas. Many approaches taking advantage of multiple antennas exist, depending on the criteria leading to their elaboration: maximum received SNR, maximum diversity, maximum bit-rate increase, and each of them having a different implementation complexity. The objective of the research here is to investigate different methods, and to compare them in terms of overall cell throughput increase. This enables to assess in a fair way the improvement brought by multiple antennas to the overall system. Intra-system reconfigurability of the devices will also be investigated, so as to adapt to varying propagation conditions, and this can also be set in a software definable radio perspective, in which different declinations with varying complexity are obtained from a given approach. Finally, the algorithmic studies are completed with real-time demonstration of the concepts, enabled by a fast prototyping approach on a research prototype, so that implementability of the designs is ensured as soon as possible. It is also expected that in return, real-time experiments will enrich the algorithm design.

1 Wireless World Research Forum (WWRF)

.2 State of the art in the area (including important references)

The use of multiple antennas in wireless systems has been studied for years, first with the aim of improving the link budget, with beamforming techniques, and more recently, to take advantage of the spatial diversity within space-time coding techniques. Such techniques rely or not on the processing of parallel data flows, which enable to increase the bit-rate as an alternative to a modulation order increase. The combination of the OFDM modulation with space-time processing techniques is rather natural, and takes advantage of the equivalent channel seen after OFDM demodulator, and composed of parallel flat fading paths. This gives an easy access to space-frequency-time diversity degrees, that need to be carefully exploited. A corresponding theoretical capacity analysis of the schemes can be found in 1. After the first general studies developed for flat fading channels in 3 and 4, space-time block coding has been applied in OFDM systems, as in 5. In 6, the problem of generalizing the orthogonal designs to the 4 transmit antenna case is adressed, and it is shown that it is possible to combine the advantages of orthogonality, rate 1 and full diversity in a space-time block code with some prior knowledge of the transmission channel. In parallel, space-time trellis codes have been designed and described in 7. Their application to OFDM has been studied in 8, revisiting optimization criteria in the OFDM case, and showing that the concatenation of a convolutional code (outer code) with a space-time trellis code (inner code) can be outperformed by a space-time trellis code designed with embedded temporal diversity. All these studies apply the space-time codes to the OFDM modulation on a subcarrier basis. Recently, some new designs have appeared, such as in 9, in which the coding is done in the 3 dimensions of space, time and frequency simultaneously. However, such designs have not been extended to a large number of subcarriers so far. Among other approaches that have been applied to OFDM, it is worth mentioning receive diversity, discussed for instance in 10, and in 11. In addition, the knowledge of the channel at the emitter open the path to many other methods, such as the one described in 12, based on the Singular Value Decomposition of the channel matrix. In general, coherent detection requires to perform an estimation of the channel parameters, and the related computational cost makes it a very sensitive topic to study carefully. Previous references have addressed this matter, such as 13, and more recently 14, and references therein.

.3 Possible approach

This section is devoted to the application of space-time block coded OFDM in the context of 5GHz WLANs, and shows performance enhancements obtained with 2 and 4 transmit antennas, combined with 1, 2 and 4 receive antennas. In a next step, these results will be compared with the ones obtained for competitive techniques,

2 Wireless World Research Forum (WWRF) first in terms of performance, then in terms of cell throughput, for the same peak data rate.

.A Space-Time Block Coding in OFDM-based WLANs

In Figure 1 and 2, the Packet Error Rate (PER) of the systems is plotted vs the SNR received per antenna. The WLAN systems are based on the IEEE80.211a/HIPERLAN2 Physical Layer, on top of which a space-time block coding stage per subcarrier is added. The target bit-rate is 12Mbps with 2x54 bytes packets, and the channel model used is BRAN A. In Figure 1, the space-time block code used is based on the Alamouti technique 3, using 2 transmit antennas, and the improvement from 1X1 to 2X2 antennas is of 8.7dB. In Figure 2, it is based on the solutions presented in 15 (STTD-OTD), in 161718 (ABBA code) and in 4 (Tarokh code), based on 4 Tx antennas. Extending the 2X2 configuration to a 4X4 one gives an additional 4.2dB gain.

.B Cell throughput increase obtained with space-time block coded OFDM

Figure 3 is based on a throughput analysis in a static environment. It takes into account the PER results over frequency selective fast fading, path loss, shadowing. IT can be seen on this curve that the 8.7dB PER improvement obtained with the 2x2 space-time block coding solution corresponds to a distance increase from 21 to 38m, which is significant.

Figure 1: Space-Time block Coded OFDM system with 2Tx antennas

Figure 2: Space-Time block Coded OFDM system with 4Tx antennas

Figure 3: Throughput vs distance between AP and MT

.C Channel estimation results In parallel, Least Squares (LS) and Minimum Mean Square Error (MMSE) estimators for the MTMR channel have been studied, both in the Frequency Domain (FD) and in the Time Domain (TD), for the same context. PER performance vs SNR

4 Wireless World Research Forum (WWRF) have been obtained (see Figure 4) and show that the MMSE outperforms the LS estimator by 1dB at PER= 10-2, whereas simplified projections of the LS estimator in the time doamin approch very closely the full TD LS estimator. Please refer to 14 for more details.

Figure 4: Space-Time Block Coded OFDM system performance with 2Tx antennas and channel estimation

.4 Expected results and time frame

Future generations of WLANs will embed without doubt multiple antennas at emitter and receiver. The first expected results will show the implementability of robust space-time block coding solutions, and will demonstrate their “real-life” real- time performance within a research testbed, within one year. In parallel, a study of the performance improvement impact at the cell level will be conducted. Note that part of this work is performed in the framework of the IST project FITNESS. Further on, other MTMR methods will be considered in a similar way for final comparisons.

.5 List of References

[1] H. Bölcskei, D. Gesbert, and A. Paulraj, “On the capacity of OFDM-based multi- antenna systems,” in Proc. Int. Conf. on Acoustics, Speech, and Signal Processing, Istanbul, Turkey, June 2000. [2] H. Bölcskei, D. Gesbert, and A. Paulraj, “On the capacity of OFDM-based spatial multiplexing systems”, IEEE Trans. on Communications, Vol. 50, n°2, Feb. 2002 [3] S. M. Alamouti, “A simple transmit diversity technique for wireless communications,” IEEE J. on Sel. Areas in Comm., pp. 1451-1458, Oct. 1998. [4] V. Tarokh, H. Jafarkhani, A. R. Calderbank: “Space-Time Block Coding for Wireless Communication : Performance Results”, IEEE JSAC, vol. 17, no. 3, pp.

451-460, Mar. 1999 [5] D. Agrawal, V. Tarokh, A. Naguib, N. Seshadri, “Space-time coded OFDM for high data-rate wireless communication over wideband channels”, Proc. IEEE Vehicular Technology Conference, 1998, Vol. 3 , pp. 2232 –2236 [6] S. Rouquette, S. Merigeault, K. Gosse, “Orthogonal full diversity space-time block coding based on transmit channel state information for 4 Tx antennas “ Proc. IEEE International Conference on Communications, 2002, Vol. 1 , pp. 558 –562. [7] V. Tarokh, N. Seshadri, A.R. Calderbank, “Space-Time Codes for High Data Rate Wireless Communication: Performance Criterion and Code Construction”, IEEE Trans. On Information Theory, Vol. 44, n°2, March 1998, pp. 744-765. [8] S. Rouquette, K. Gosse , « Space-Time Coding Options for OFDM-based WLANs », Proc. IEEE Vehicular Technology Conference, Spring 2002. [9] H. Bölcskei and A. Paulraj, “Space-frequency codes for broadband fading channels”, Proc. IEEE Int. Symposium on Information Theory, 2001, p. 219. [10] A.A. Hutter, J.S. Hammerschmidt, E. de Carvalho, J.M. Cioffi, “Receive diversity for mobile OFDM systems“, IEEE Wireless Communications and Networking Conference, 2000, Vol. 2 , pp. 707 –712. [11] D. Bartolomé, X. Mestre, A. Perez-Neira, « Single Input, Multiple Output Techniques for HIPERLAN/2 », IST Summit 2001, Barcelona [12] G. G. Raleigh, J. M. Cioffi, “Spatio-Temporal Coding for Wireless Communication,” IEEE Trans. on Communications, vol. 46, pp. 357-366, March 1998. [13] Y. Li, N. Seshadri, S. Ariyavisitakul, « Channel Estimation for OFDM with Transmitter Diversity in Mobile Wireless Channels”, IEEE Journal on Selected Areas in Communications, vol. 17, n°3, March 1999. [14] A. Ribeiro Dias, S. Rouquette, K. Gosse, « MTMR Channel Estimation and Pilot Design in the Context of Space-Time Block Coded OFDM-based WLANs », IST Summit 2002, Thessaloniki, June 16-19. [15] L.M.A. Jalloul, K. Rohani, K. Kuchi, J. Chen “Performance Analysis of CDMA Transmit Diversity Methods”, in Proc. IEEE Vehicular Technology Conference, 1999, vol. 3, pp. 1326-1330. [16] O. Tirkkonen, A. Boariu, A. Hottinen, “Minimal Non-Orthogonality Rate 1 Space- Time Block Code for 3+ Tx”, Proc. Int. Symposium on Spread Spectrum Techniques and Applications, pp. 429-432, Sept. 2000. [17] H. Jafarkhani, “A Quasi-Orthogonal Space-Time Block Code”, IEEE Trans. On Communications, vol. 49, n°1, Jan. 2001. [18] C.B. Papadias, G.J. Foschini, “A Space-Time Coding Approach for Systems Employing Four Transmit Antennas”, Proc. ICASSP, Salt Lake City, June 2001.