Multi-user multiple-antenna system

Abstract

To provide a multi-user multiple-antenna system (MU-MAS).SOLUTION: A multi-user multiple-antenna system (MU-MAS) comprises: a plurality of user devices each having at least one antenna; and one or more wireless transceivers creating a plurality of concurrent and independent wireless links with the plurality of user devices, where at least one of the wireless transceiver stations has far more antennas than at least one user device. The MU-MAS computes one or more quality metrics and determines one or more MU-MAS transmission modes used over the wireless links.SELECTED DRAWING: Figure 21

Description

[Technical Field]

[0001] Priority claims This application is a continuation-in-part of U.S. patent application Ser. No. 10 / 902,978, filed Jul. 30, 2004.

[0002] The present invention relates generally to the field of communication systems, and more particularly to a system and method for distributed input / distributed output wireless communication using space-time coding techniques. [Background technology]

[0003] Space-time coding of communication signals A relatively new development in wireless technology is known as spatial multiplexing and space-time coding. One particular form of space-time coding is called MIMO, for "multiple input multiple output," because several antennas are used at each end. By using multiple antennas for transmission and reception, multiple independent radio waves can be transmitted simultaneously within the same frequency range. The following paper provides an overview of MIMO:

[0004] David Gesbert, Member, IEEE; Mansoor Shafi, Fellow, IEEE; Da-shan Shiu, Member, IEEE; Peter J. Smith, Member, IEEE; and Ayman Naguib, Senior Member, IEEE, "From Theory to Implementation: An Overview of MIMO Space-Time Coded Wireless Systems," Journal of IEEE Society on Selected Areas of Communications, Vol. 21, No. 3, April 2003.

[0005] David Gesbert, IEEE Member, Helmut Bolcskei, IEEE Member, Dhananjay A. Gore, and Arogyaswami J. Paulraj, IEEE Fellow, "Outdoor MIMO Wireless Channels: Models and Performance Predictions," IEEE Transactions on Communications, Vol. 50, No. 12, December 2002.

[0006] Essentially, MIMO technology is based on the use of spatially distributed antennas to create parallel spatial data streams within a common frequency band. Radio waves are transmitted in such a way that, even though they are transmitted within the same frequency band, the individual signals can be separated and demodulated at the receiver, resulting in multiple statistically independent (i.e., virtually separate) communication channels. That is, in contrast to standard wireless communication systems that attempt to reject multipath signals (i.e., multiple signals of the same frequency delayed in time and modified in amplitude and phase), MIMO can achieve higher throughput and improved signal-to-noise ratio within a given frequency band using uncorrelated or weakly correlated multipath signals. As an example, MIMO technology achieves much higher throughput under similar power and signal-to-noise ratio (SNR) conditions where conventional non-MIMO systems can only achieve lower throughput. This feature allows http: / / www.cdmatech.com / products / what mimo deliveries.isp This is explained on the website of Qualcomm Incorporated (Qualcomm is one of the largest providers of wireless technology) in a page titled "What MIMO Delivers" at: http: / / www.qualcomm.com / (Qualcomm is one of the largest providers of wireless technology). MIMO is the only multiple antenna technology that increases spectrum capability by delivering double or more of the system's peak data rate per channel or per MHz of spectrum. More specifically, for wireless LAN or Wi-Fi® applications, QUALCOMM's fourth generation MIMO technology delivers speeds of 315 Mbps, or 8.8 Mbps / MHz, in 36 MHz of spectrum. Compare this to the peak capability of 802.11a / g (even with beamforming or diversity techniques), which delivers only 54 Mbps, or 3.18 Mbps / MHz, in 17 MHz of spectrum.

[0007] MIMO systems typically face a practical limit of less than 10 antennas per device (and therefore less than 10X throughput improvement in the network) for several reasons. 1. Physical limitations: There must be sufficient separation between the MIMO antennas on a given device so that each receives a statistically independent signal. While MIMO throughput improvements can be seen with antenna spacings even a small fraction of a wavelength, efficiency deteriorates rapidly when antennas are closely spaced, resulting in a lower MIMO throughput multiplier.

[0008] See, for example, the following references: [1] D.S. Shiu, G.J. Foschini, M.J. Gans, and J.M. Khan, "Fading Correlation and Its Impact on the Performance of Multi-Element Antenna Systems," IEEE Transactions on Communications, Vol. 48, No. 3, pp. 502-513, March 2000. [2] V. Pohl, V. Jungnickel, T. Haustein, and Helmoldt, "Antenna Spacing in MIMO Indoor Channels," Proceedings of the IEEE Conference on Veh. Technol. conf., Vol. 2, pp. 749-753, May 2002. [3] M. Stoytchev, H. Safar, A. L. Moustakas, and S. Simon, "Small Antenna Arrays for MIMO Applications," Antennas and Prop. Symp., IEEE Proceedings of the IEEE Conference on Signal Processing, Vol. 3, pp. 708-711, July 2001. [4] A. Forenza and R.W. Heath, Jr., "Effects of Antenna Geometry on MIMO Communications in Indoor Clustered Channels," Antennas and Prop. Symp., IEEE Proceedings, Vol. 2, pp. 1700-1703, June 2004.

[0009] Also, for small antenna spacing, the performance of MIMO systems can be degraded due to the effects of mutual coupling.

[0010] See, for example, the following references: [5] M.J. Fakhereddin and K.R. Dandekar, “Combined Impact of Polarization Diversity and Mutual Coupling on MIMO Performance,” Antennas and Prop. Symp., IEEE Proceedings of the IEEE Conference on Antennas and Prop. Symp., Vol. 2, pp. 495-498, June 2003. [7] P.N. Letcher, M. Dean, and A.R.Nix, "Mutual Coupling in Multielement Array Antennas and Its Impact on MIMO Channel Performance," IEEE Electronics Letters, Vol. 39, pp. 342-344, February 2003. [8] V. Jungnickel, V. Pohl, and C. Von Helmolt, "Performance of MIMO Systems with Closely Spaced Antennas," IEEE Communications Letters, Vol. 7, pp. 361-363, August 2003.

[10] J.W. Wallace and M.A. Jensen, "Termination-Dependent Diversity Performance of Coupled Antennas: A Network-Theoretic Analysis," IEEE Transactions on Antenna Propagation, Vol. 52, pp. 98-105, January 2004.

[13] C. Waldschmidt, S. Schulteis, and W. Wiesbeck, "A complete RF system model for the analysis of small MIMO arrays," IEEE Transactions on Veh. Technol., Vol. 53, pp. 579-586, May 2004.

[14] M.L. Morris and M.A. Jensen, "Network Models for MIMO Systems with Coupled Antennas and Noisy Amplifiers," IEEE Transactions on Antenna Propagation, Vol. 53, pp. 545-552, January 2005.

[0011] Furthermore, because the antennas are closely spaced together, they generally must be made smaller, which can also affect antenna efficiency.

[0012] See, for example, the following references:

[15] H.A. Wheeler, "Small Antennas," IEEE Transactions on Antenna Propagation, Vol. AP-23, No. 4, pp. 462-469, July 1975.

[16] J.S.M. Clean, "A Review of Fundamental Limits on the Radiation Q of Electrically Small Antennas," IEEE Transactions on Antenna Propagation, Vol. 44, No. 5, pp. 672-676, May 1996.

[0013] Finally, at lower frequencies and longer wavelengths, the physical size of a single MIMO device can become unwieldy. An extreme example is in the HF band, where MIMO device antennas should be separated from each other by 10 meters or more. 2. Noise Limitation. Each MIMO receiver / transmitter subsystem generates a certain level of noise. As more of these subsystems are placed in close proximity to each other, the noise level increases. Meanwhile, the need to distinguish between an ever-increasing number of individual signals in a multi-antenna MIMO system requires increasingly lower noise levels. 3. Cost and Power Constraints. While there are MIMO applications where cost and power consumption are not an issue, for typical wireless products, both cost and power consumption are critical constraints on successful product development. A separate RF subsystem, including separate analog-to-digital (A / D) and digital-to-analog (D / A) converters, is required for each MIMO antenna. Unlike many aspects of digital systems that respond to Moore's Law (an empirical observation by Intel co-founder Gordon Moore that the number of transistors on an integrated circuit for the smallest component doubles its cost every 24 months, source: http: / / www.intel.com / technology / mooreslaw / ), such analog-intensive subsystems typically have a fixed physical size and power requirements, and scale linearly in cost and power. Therefore, multi-antenna MIMO devices are expected to be prohibitively expensive and power-hungry compared to single-antenna devices.

[0014] As a result of the above, most MIMO systems considered today are on the order of two to four antennas, providing a two- to four-fold increase in throughput and some increase in SNR due to the diversity benefits of the multiple antenna system. While MIMO systems with up to ten antennas have been considered (especially at higher microwave frequencies due to shorter wavelengths and closer antenna spacing), numbers significantly greater than that are impractical for all but very specialized and cost-intensive applications.

[0015] Virtual Antenna Array One special application of MIMO-type techniques is the virtual antenna array. Such a system is proposed in a research paper, "Steps Towards MIMO: Virtual Antenna Arrays," by Mischa Dohler and Hamid Aghvami, presented at the European Cooperation in Scientific and Technological Research (EURO-COST), Barcelona, ​​Spain, January 15-17, 2003, King's College Telecommunications Research Centre, London, UK.

[0016] A virtual antenna array, as presented in this paper, is a system of cooperating wireless devices (such as cell phones) that communicate among themselves on a communication channel separate from their primary communication channel to their serving base station, if and when they are close enough to each other (e.g., if they are GSM cell phones in the UHF band, this could be the 5 GHz Industrial, Scientific, and Medical (ISM) radio band), so that they work in cooperation. This allows, for example, a single-antenna device to relay information between several devices that are within range of each other (in addition to being within range of the base station) and potentially achieve MIMO-like throughput increases by operating as if they were one physical device with multiple antennas. [Prior art documents] [Patent documents]

[0017] [Patent Document 1] U.S. Patent Application Serial No. 10 / 902,978 [Patent Document 2] U.S. Patent Application Serial No. 10 / 817,731 [Non-patent literature]

[0018] [Non-Patent Document 1] David Gesbert, Member, IEEE; Mansoor Shafi, Fellow, IEEE; Da-shan Shiu, Member, IEEE; Peter J. Smith, Member, IEEE; and Ayman Naguib, Senior Member, IEEE, "From Theory to Implementation: An Overview of MIMO Space-Time Coded Wireless Systems," Journal of IEEE Society on Selected Areas of Communications, Vol. 21, No. 3, April 2003. [Non-patent document 2] David Gesbert, IEEE Member, Helmut Bolcskei, IEEE Member, Dhananjay A. Gore, and Arogyaswami J. Paulraj, IEEE Fellow, "Outdoor MIMO Wireless Channels: Models and Performance Predictions," IEEE Transactions on Communications, Vol. 50, No. 12, December 2002. [Non-patent document 3] D.S. Shiu, G.J. Foschini, M.J. Gans, and J.M. Khan, "Fading Correlation and Its Impact on the Performance of Multi-Element Antenna Systems," IEEE Transactions on Communications, Vol. 48, No. 3, pp. 502-513, March 2000. [Non-patent document 4] V. Pohl, V. Jungnickel, T. Haustein, and Helmoldt, "Antenna Spacing in MIMO Indoor Channels," Proceedings of the IEEE Conference on Veh. Technol. conf., Vol. 2, pp. 749-753, May 2002. [Non-Patent Document 5] M. Stoytchev, H. Safar, A. L. Moustakas, and S. Simon, "Small Antenna Arrays for MIMO Applications," Antennas and Prop. Symp., IEEE Proceedings, Vol. 3, pp. 708-711, July 2001. [Non-patent document 6] A. Forenza and R.W. Heath, Jr., "Effects of Antenna Geometry on MIMO Communications in Indoor Clustered Channels," Antennas and Prop. Symp., IEEE Proceedings, Vol. 2, pp. 1700-1703, June 2004. [Non-Patent Document 7] MJ Fakhereddin and KR Dandekar, "Combined Impact of Polarization Diversity and Mutual Coupling on MIMO Performance," Antennas and Prop. Symp., IEEE Proceedings, Vol. 2, pp. 495-498, June 2003. [Non-patent document 8] P.N. Letcher, M. Dean, and A.R.Nix, "Mutual Coupling in Multielement Array Antennas and Its Impact on MIMO Channel Performance," IEEE Electronics Letters, Vol. 39, pp. 342-344, February 2003. [Non-Patent Document 9] V. Jungnickel, V. Pohl, and C. Von Helmolt, "Performance of MIMO Systems with Closely Spaced Antennas," IEEE Communications Letters, Vol. 7, pp. 361-363, August 2003. [Non-Patent Document 10] JW Wallace and M.A. Jensen, "Termination-Dependent Diversity Performance of Coupled Antennas: A Network-Theoretical Analysis," IEEE Transactions on Antenna Propagation, Vol. 52, pp. 98-105, January 2004. [Non-Patent Document 11] C. Waldschmidt, S. Schulteis, and W. Wiesbeck, "A Complete RF System Model for the Analysis of Small MIMO Arrays," IEEE Transactions on Veh. Technol., Vol. 53, pp. 579-586, May 2004. [Non-Patent Document 12] ML Morris and MA Jensen, "Network Models for MIMO Systems with Coupled Antennas and Noisy Amplifiers," IEEE Transactions on Antenna Propagation, Vol. 53, pp. 545-552, January 2005. [Non-Patent Document 13] H.A. Wheeler, "Small Antennas," IEEE Transactions on Antenna Propagation, Vol. AP-23, No. 4, pp. 462-469, July 1975. [Non-Patent Document 14] J.S.M. Clean, "A Review of Fundamental Limits on the Radiation Q of Electrically Small Antennas," IEEE Transactions on Antenna Propagation, Vol. 44, No. 5, pp. 672-676, May 1996. [Non-Patent Document 15] Mischa Dohler and Hamid Aghvami, "Steps Towards MIMO: Virtual Antenna Arrays," King's College Telecommunications Research Centre, London, UK, January 15-17, 2003, European Cooperation in Scientific and Technological Research (EURO-COST), Barcelona, ​​Spain. [Non-Patent Document 16] L. Dong, H. Ling, and R.W. Heath, Jr., "Multiple-Input Multiple-Output Wireless Communication Systems Using Antenna Pattern Diversity," Proceedings of the IEEE Global Telecom Conf., Vol. 1, pp. 997-1001, November 2002. [Non-Patent Document 17] R. Vaughan, "Switched Parasitic Elements for Antenna Diversity," IEEE Transactions on Antenna Propagation, Vol. 47, pp. 399-405, February 1999. [Non-Patent Document 18] P. Mattheijssen, M. H. A. Herben, G. Dolmans, and L. Leyten, "Antenna Pattern Diversity Versus Spatial Diversity for Handheld Use," IEEE Transactions on Veh. Technol., Vol. 53, pp. 1035-1042, July 2004. [Non-Patent Document 19] C.B. Dietrich Jr., K. Dietze, J.R. Nealy, and W.L. Stutzman, "Spatial, Polarization, and Pattern Diversity for Wireless Handheld Terminals," Proceedings of the IEEE Antennas and Propagation Symposium, Vol. 49, pp. 1271-1281, September 2001. [Non-Patent Document 20] A. Forenza and R.W. Heath, Jr., "Pattern Diversity Benefits of a Two-Element Array of Circular Patch Antennas in Indoor Clustered MIMO Channels," IEEE Transactions on Communications, Vol. 54, No. 5, pp. 943-954, May 2006. [Non-Patent Document 21] M.D. Benedetto and P. Mandarini, "Analysis of the Impact of I / Q Baseband Filter Mismatch in OFDM Modems," Wireless Personal Communications, pp. 175-186, 2000. [Non-Patent Document 22] S. Schuchert and R. Hasholzner, "A Novel I / Q Imbalance Correction Scheme for Receiving OFDM Signals," IEEE Transactions on Consumer Electronics, August 2001. [Non-Patent Document 23] M. Valkama, M. Renfors, and V. Koivunen, "Advanced Methods for I / Q Imbalance Correction in Communications Receivers," IEEE Transactions on Signal Processing, October 2001. [Non-Patent Document 24] R. Rao and B. Daneshrad, "Analysis of I / Q Mismatch and Cancellation Methods for OFDM Systems," IST Mobile Communications Summit, June 2004. [Non-Patent Document 25] A. Tarighat, R. Bagheri, and A. H. Sayed, "Compensation Scheme and Performance Analysis of IQ Imbalance in OFDM Receivers," IEEE Transactions on Signal Processing, Vol. 53, pp. 3257-3268, August 2005. [Non-Patent Document 26] R. Rao and B. Daneshrad, "I / Q Mismatch Cancellation in MIMO-OFDM Systems," in Personal, Indoor, and Mobile Radio Communications, 2004, PIMRC2004, 15th IEEE International Symposium, Vol. 4, 2004, pp. 2710-2714. [Non-Patent Document 27] R.M. Rao, W. Zhu, S. Lang, C. Oberli, D. Browne, J. Bhatia, J.F. Rigon, J. Wang, P. Gupta, H. Lee, D.N. Liu, S.G. Wong, M. Fitz, B. Daneshrad, and O. Takeshita, "A Multiple Antenna Testbed for Wireless Communications Research and Education," IEEE Communications Journal, Vol. 42, No. 12, pp. 72-81, December 2004. [Non-patent document 28] S. Lang, M.R. Rao, and B. Daneshrad, "Design and Development of a 5.25GHz Software-Defined Radio OFDM Communications Platform," IEEE Communications Journal (IEEE), Vol. 42, No. 6, pp. 6-12, June 2004. [Non-Patent Document 29] A. Tarighat and A. H. Sayed, "MIMO-OFDM Receivers for Systems with IQ Imbalance," IEEE Transactions on Signal Processing, Vol. 53, pp. 3583-3596, September 2005. [Non-Patent Document 30] Q. H. Spencer, A. L. Windlehurst, and M. Haardt, "Zero-Forcing Methods for Downlink Spatial Multiplexing in Multiuser MIMO Channels," IEEE Transactions on Signal Processing, Vol. 52, pp. 461-471, February 2004. [Non-Patent Document 31] K.K. Wong, R.D. Murch, and K.B. Letaief, "Joint Channel Diagonalization for Multiuser MIMO Antenna Systems," IEEE Transactions on Wireless Communications, Vol. 2, pp. 773-786, July 2003. [Non-Patent Document 32] LU Choi and R.D. Murch, "Transmit Preprocessing Techniques for Multiuser MIMO Systems Using Decomposition Techniques," IEEE Transactions on Wireless Communications, Vol. 3, pp. 20-24, January 2004. [Non-Patent Document 33] Z. Shen, J. G. Andrews, R. W. Heath, and B. L. Vans, "A Low-Complexity User Selection Algorithm for Multiuser MIMO Systems with Block Diagonalization," IEEE Transactions on Signal Processing, accepted for publication, September 2005. [Non-Patent Document 34] Z. Shen, R. Chen, J. G. Andrews, R. W. Heath, and B. L. Vans, "Sum Function of Multiuser MIMO Broadcast Channels with Block Diagonalization," submitted to the IEEE Transactions on Wireless Communications, October 2005. [Non-Patent Document 35] R. Chen, R.W. Heath, and J.G. Andrews, "Transmit Selection Diversity for Single-Precoded Multiuser Spatial Multiplexing Systems with Linear Receivers," accepted for IEEE Transactions on Signal Processing, 2005. [Non-Patent Document 36] A. Tarighat and A. H. Sayed, "MIMO-OFDM Receivers for Systems with IQ Imbalance," IEEE Transactions on Signal Processing, Vol. 53, pp. 3583-3596, September 2005. [Non-Patent Document 37] A. Tarighat, R. Bagheri, and A. H. Sayed, "Compensation Scheme and Performance Analysis of IQ Imbalance in OFDM Receivers," IEEE Transactions on Signal Processing, Vol. 53, pp. 3257-3268, August 2005. [Non-Patent Document 38] V. Tarokh, H. Jafarkhani, and A. C. Calderbank, "Space-Time Block Codes Using Orthogonal Coordination," IEEE Transactions on Information Theory, Vol. 45, pp. 1456-467, July 1999. [Non-Patent Document 39] R.W. Seath, Jr., S. Sandhu, and A.J. Paulraj, "Antenna Selection for Spatial Multiplexing Systems with Linear Receivers," IEEE Transactions on Communications, Vol. 5, pp. 142-144, April 2001. [Non-Patent Document 40] GJ Foschini, GD Golden, RA Valenzuela, and PW Wolniansky, "Simplified Processing for High Spectral Efficiency Wireless Communications Using Multi-Element Arrays," Proceedings of the IEEE Conference on Selected Areas of Communications, Vol. 17, No. 11, pp. 1841-1852, November 1999. [Non-Patent Document 41] L. Zheng and D.N. C. Tse, "Diversity and Multiplexing: Fundamental Tradeoffs in Multi-Antenna Channels," IEEE Transactions on Information Theory, Vol. 49, No. 5, pp. 1073-1096, May 2003. [Non-Patent Document 42] R.W. Seath, Jr., S. Sandhu, and A.J. Paulraj, "Switching Between Diversity and Multiplexing in MIMO Systems," IEEE Transactions on Communications, Vol. 53, No. 6, pp. 962-968, June 2005. [Non-Patent Document 43] S. Catreux, V. Erceg, D. Gesbert, and R.W. Heath, Jr., "Adaptive Modulation and MIMO Coding for Broadband Wireless Data Communication Networks," IEEE Communications Journal, Vol. 2, pp. 108-115, June 2002. [Non-Patent Document 44] M.R. McKay, I.B. Collings, A. Forenza, and R.W. Heath, Jr., "Coded MIMO Multiplexing / Beamforming Switching in Spatially Correlated Rayleigh Channels," accepted for publication in the IEEE Transactions on Veh. Tech., December 2007. [Non-Patent Document 45] A. Forenza, M. R. McKay, R. W. Heath Jr., and I. B. Collings, "Switching between OSTBC and Spatial Multiplexing with a Linear Receiver in Spatially Correlated MIMO Channels," Proceedings of the IEEE Veh. Technol. Conf., Vol. 3, pp. 1387-1391, May 2006. [Non-Patent Document 46] M.R. McKay, I.B. Collings, A. Forenza, and R.W. Heath, Jr., "Throughput-Based Adaptive MIMO-BICM Techniques for Spatially Correlated Channels," to be published in the Proceedings of the IEEE-ICC Conference, June 2006. [Non-Patent Document 47] R.W. Seath, Jr., S. Sandhu, and A.J. Paulraj, "Switching Between Diversity and Multiplexing in MIMO Systems," IEEE Transactions on Communications, Vol. 53, No. 6, pp. 962-968, June 2005. [Non-Patent Document 48] S. Catreux, V. Erceg, D. Gesbert, and R.W. Heath, Jr., "Adaptive Modulation and MIMO Coding for Broadband Wireless Data Communication Networks," IEEE Communications Journal, Vol. 2, pp. 108-115, June 2002. [Non-Patent Document 49] A. Forenza, M.R. McKay, A. Pandharipande, R.W. Heath Jr., and I.B. Collings, "Adaptive MIMO Transmission Exploiting Spatially Correlated Channel Features," IEEE Transactions on Veh. Tech., Vol. 56, No. 2, pp. 619-630, March 2007. [Non-Patent Document 50] M.R. McKay, I.B. Collings, A. Forenza, and R.M. McKay, I.B. Collings, A. Forenza, and R.W. Heath, Jr., "Throughput-Based Adaptive MIMO-BICM Techniques for Spatially Correlated Channels," to be published in the Proceedings of the IEEE-ICC Conference, June 2006. [Non-Patent Document 51] M. Sharif and B. Hassibi, "On the Performance of MIMO Broadcast Channels with Partial Side Information," IEEE Transactions on Information Theory, Vol. 51, pp. 506-522, February 2005. [Non-Patent Document 52] W. Choi, A. Forenza, J.G. Andrews, and R.W. Heath, Jr., "Opportunistic Space Division Multiple Access Schemes with Beam Selection," to be published in IEEE Transactions on Communications. [Non-Patent Document 53] X. Zhuang, FW Vook, KL Baum, TA Thomas, and M. Cudak, "Channel Models for Link and System Level Simulation," IEEE 802.16 Broadband Wireless Access Working Group, September 2004. [Non-Patent Document 54] T. Fusco and M. Tanda, "Insensitive Frequency Offset Estimation for OFDM / OQAM Systems," IEEE Transactions on Signal Processing, Vol. 55, pp. 1828-1838, 2007. [Non-Patent Document 55] E. Serpedin, A. Chevreuil, G.B. Giannakis, and P. Loubaton, "Blind Channel and Carrier Frequency Offset Estimation Using Periodic Modulation Precoders," IEEE Transactions on Signal Processing, Vol. 48, No. 8, pp. 2389-2405, August 2000. [Non-Patent Document 56] JJ van de Beek, M. Sandell, and P.O. Borjesson, "ML Estimation of Time and Frequency Offsets in OFDM Systems," IEEE Transactions on Signal Processing, Vol. 45, No. 7, pp. 1800-1805, July 1997. [Non-Patent Document 57] U. Tureli, H. Liu, and M.D. Zoltowski, "OFDM blind carrier offset estimation: ESPRIT," IEEE Transactions on Communications, Vol. 48, No. 9, pp. 1459-1461, September 2000. [Non-Patent Document 58] M. Luise, M. Marselli, and R. Reggiannini, "Low-Complexity Insensitive Carrier Frequency Recovery for OFDM Signals over Frequency-Selective Wireless Channels," IEEE Transactions on Communications, Vol. 50, No. 7, pp. 1182-1188, July 2002. [Non-Patent Document 59] P.H. Moose, "Orthogonal Frequency Division Multiplexing Frequency Offset Compensation Method," IEEE Transactions on Communications, Vol. 42, No. 10, pp. 2908-2914, October 1994. [Non-Patent Document 60] T.M. Schmidl and D.C. Cox, "Robust Frequency and Timing Synchronization for OFDM," IEEE Transactions on Communications, Vol. 45, No. 12, pp. 1613-1621, December 1997. [Non-Patent Document 61] M. Luise, M. Marselli, and R. Reggiannini, "Carrier Frequency Acquisition and Tracking for OFDM Systems," IEEE Transactions on Communications, Vol. 44, No. 11, pp. 1590-1598, November 1996. [Non-Patent Document 62] JJ van de Beek, M. Sandell, and P.O. Borjesson, "ML Estimation of Time and Frequency Offsets in OFDM Systems," IEEE Transactions on Signal Processing, Vol. 45, No. 7, pp. 1800-1805, July 1997. [Non-Patent Document 63] K. Lee and J. Chun, "Frequency Offset Estimation for MIMO and OFDM Systems Using Orthogonal Training Sequences," IEEE Transactions on Veh. Technol., Vol. 56, No. 1, pp. 146-156, January 2007. [Non-Patent Document 64] M. Ghogho and A. Swami, "Training Design for Multipath Channel and Frequency Offset Estimation in MIMO Systems," IEEE Transactions on Signal Processing, Vol. 54, No. 10, pp. 3957-3965, October 2005. [Non-Patent Document 65] C. Oberli and B. Daneshrad, "Maximum Likelihood Tracking Algorithm for MIMO OFDM," in 2004 IEEE International Conference on, Vol. 4, June 4-24, 2004, pp. 2468-2472. [Non-Patent Document 66] A. Kannan, T. P. Krauss, and M. D. Zoltowski, "Separation of Co-Channel Signals Under Imperfect Timing and Carrier Synchronization," IEEE Transactions on Veh. Technol., Vol. 50, No. 1, pp. 79-96, January 2001. [Non-Patent Document 67] T. Tang and R.W. Heath, "Joint Frequency Offset Estimation and Interference Cancellation for MIMO-OFDM Systems [Mobile Radio]," in VTC 2004 Fall, 2004 IEEE 60th Vehicle Technology Conference, Vol. 3, pp. 1553-1557, September 26-29, 2004. [Non-Patent Document 68] X. Dai, "Carrier Frequency Offset Estimation for OFDM / SDMA Systems Using Continuous Pilots," IEEE Transactions on Communications, Vol. 152, pp. 624-632, October 7, 2005. [Non-Patent Document 69] L. Haring, S. Bieder, and A. Czylwik, "Residual Carrier and Sampling Frequency Synchronization in Multiuser OFDM Systems," in 2006 VTC, Spring 2006, IEEE 63rd Vehicle Technology Conference, Vol. 4, pp. 1937-1941, 2006. [Non-Patent Document 70] O. Besson and P. Stoica, "On Parameter Estimation for MIMO Flat Fading Channels with Frequency Offsets," IEEE Transactions on Signal Processing, Vol. 51, No. 3, pp. 602-613, March 2003. U.S. Patent Application No. 20060023803 [Non-Patent Document 71] LU Choi and R.D. Murch, "Transmit Preprocessing Methods for Multiuser MIMO Systems Using Decomposition Techniques," IEEE Transactions on Wireless Communications, Vol. 3, pp. 20-24, January 2004. [Non-Patent Document 72] P.H. Moose, "Orthogonal Frequency Division Multiplexing Frequency Offset Compensation Method," IEEE Transactions on Communications, Vol. 42, No. 10, pp. 2908-2914, October 1994. [Non-Patent Document 73] A. J. Coulson, "Maximum Likelihood Synchronization of OFDM Using Pilot Symbols: An Analysis," IEEE Journal on Selected Areas in Communications, Vol. 19, No. 12, pp. 2495-2503, December 2001. [Non-Patent Document 74] A. J. Coulson, "Maximum Likelihood Synchronization of OFDM Using Pilot Symbols: An Algorithm," IEEE Journal on Selected Areas in Communications, Vol. 19, No. 12, pp. 2486-2494, December 2001. [Non-Patent Document 75] H. Minn, V. K. Bhargava, and K. B. Letaief, "Robust Timing and Frequency Synchronization for OFDM Systems," IEEE Transactions on Wireless Communications, Vol. 2, No. 4, pp. 822-839, July 2003. [Non-Patent Document 76] K. Shi and E. Serpedin, "Coarse Frame and Carrier Synchronization for OFDM Systems: New Metrics and Comparisons," IEEE Transactions on Wireless Communications, Vol. 3, No. 4, pp. 1271-1284, July 2004. [Non-Patent Document 77] M. Morelli, A.D. Andrea, and U. Mengali, "Resolving Frequency Ambiguity in OFDM Systems," IEEE Communications Letters, Vol. 4, No. 4, pp. 134-136, April 2000. [Non-Patent Document 78] M. Morelli and U. Mengali, "An Improved Frequency Offset Estimator for OFDM Applications," IEEE Communications Letters, Vol. 3, No. 3, pp. 75-77, March 1999. [Non-Patent Document 79] D. Chu, "Polyphase Codes with Good Periodic Correlation Properties (Concordance)," IEEE Transactions on Information Theory, Vol. 52, No. 4, pp. 531-532, July 1972. [Non-Patent Document 80] http: / / www.cdmatech.com / products / what mimo deliveries.isp"Qualcomm Incorporated" [Non-Patent Document 81] Moore's Law, http: / / www.intel.com / technology / mooreslaw / Summary of the Invention [Problem to be solved by the invention]

[0019] However, in practice, such systems are extremely difficult to implement and have limited usefulness. For one, there are currently at least two separate communication paths per device that must be maintained to achieve improved throughput, with the second relay link often having uncertain availability. Also, devices are more expensive, physically larger, and consume more power because they have at least a second communication subsystem and greater computational requirements. Furthermore, the system relies on a very high degree of real-time coordination of all devices, potentially across various communication links. Finally, increasing simultaneous channel utilization (e.g., simultaneous phone call transmissions using MIMO techniques) increases the computational load for each device (potentially exponentially as channel utilization increases linearly), which may be highly impractical for portable devices with strict power and size constraints. [Means for solving the problem]

[0020] A system and method for correcting frequency and phase offsets in a multiple antenna system (MAS) with multi-user (MU) transmission (MU-MAS) is described. For example, a method according to one embodiment of the present invention includes transmitting a training signal from each antenna of a base station to one or each of a plurality of wireless client devices that analyze each training signal to generate frequency offset correction data, and receiving the frequency offset correction data at the base station; calculating MU-MAS precoder weights based on the frequency offset correction data at the transmitter to pre-cancel the frequency offset; precoding the training signal using the MU-MAS precoder weights to generate a precoded training signal for each antenna of the base station; transmitting the precoded training signal from each antenna of the base station to each of a plurality of wireless client devices that each analyzes a respective training signal to generate channel characterization data, and receiving the channel characterization data at the base station; calculating, based on the channel characterization data, a plurality of MU-MAS precoder weights calculated to pre-cancel frequency and phase offset and / or inter-user interference; precoding data using the MU-MAS precoder weights to generate a precoded data signal for each antenna of the base station;

[0021] The present invention can be better understood from the following detailed description taken in conjunction with the drawings. [Brief explanation of the drawings]

[0022] [Figure 1] FIG. 1 illustrates a prior art MIMO system. [Figure 2] FIG. 1 illustrates an N-antenna base station communicating with multiple single-antenna client devices. [Figure 3] FIG. 1 illustrates a three-antenna base station communicating with three single-antenna client devices. [Figure 4] FIG. 1 illustrates a training signal method used in one embodiment of the present invention. [Figure 5] FIG. 4 illustrates channel characterization data transmitted from a client device to a base station according to one embodiment of the present invention. [Figure 6] FIG. 1 illustrates multiple input distributed output (MIDO) downstream transmission according to one embodiment of the present invention. [Figure 7] FIG. 2 illustrates multiple-input multiple-output (MIMO) upstream transmission according to one embodiment of the present invention. [Figure 8] 1 illustrates a base station cycling through different client groups to allocate throughput according to one embodiment of the present invention. [Figure 9] FIG. 1 illustrates proximity-based client grouping according to one embodiment of the present invention. [Figure 10] FIG. 1 illustrates an embodiment of the present invention employed within an NVIS system. [Figure 11] 1 illustrates an embodiment of a DIDO transmitter having an I / Q correction functional unit. [Figure 12] FIG. 1 illustrates a DIDO receiver having an I / Q correction functional unit. [Figure 13] FIG. 1 illustrates an embodiment of a DIDO-OFDM system with I / Q compensation. [Figure 14] FIG. 1 illustrates one embodiment of DIDO 2×2 performance with and without I / Q compensation. [Figure 15] FIG. 1 illustrates one embodiment of DIDO 2×2 performance with and without I / Q compensation. [Figure 16] FIG. 1 illustrates one embodiment of SER (Symbol Error Rate) with and without I / Q compensation for different QAM constellations. [Figure 17] FIG. 1 illustrates one embodiment of DIDO 2×2 performance with and without correction at different user equipment positions. [Figure 18] FIG. 1 illustrates one embodiment of SER with and without I / Q compensation in an ideal (iid (independent and identically distributed)) channel. [Figure 19] FIG. 1 illustrates one embodiment of a transmitter framework for an adaptive DIDO system. [Figure 20] FIG. 1 illustrates one embodiment of a receiver framework for an adaptive DIDO system. [Figure 21] FIG. 1 illustrates an embodiment of a method for adaptive DIDO-OFDM. [Figure 22] FIG. 1 illustrates one embodiment of an antenna arrangement for DIDO measurements. [Figure 23] FIG. 1 illustrates an embodiment of an array configuration for a mixed-order DIDO system. [Figure 24] FIG. 1 illustrates the performance of a different-order DIDO system. [Figure 25] FIG. 1 illustrates an embodiment of an antenna arrangement for DIDO measurements. [Figure 26] FIG. 1 illustrates one embodiment of DIDO 2x2 performance with 4-QAM and FEC rate ½ as a function of user equipment location. [Figure 27] FIG. 1 illustrates an embodiment of an antenna arrangement for DIDO measurements. [Figure 28] FIG. 10 illustrates how DIDO8x8 provides larger SE than DIDO2x2 for lower TX power requirements in one embodiment. [Figure 29] FIG. 1 illustrates one embodiment of DIDO 2x2 performance with antenna selection. [Figure 30] FIG. 1 illustrates the average bit error rate (BER) performance of different DIDO precoding schemes in iid channels. [Figure 31] 10 illustrates the signal-to-noise ratio (SNR) gain of ASel as a function of the number of extra transmit antennas in an iid channel. [Figure 32] FIG. 10 shows the SNR threshold as a function of the number of users (M) for block diagonalization (BD) and ASel with one and two extra antennas in iid channels. [Figure 33]FIG. 1 shows BER versus average SNR per user for two users located in the same angular direction with different values ​​of angular spread (AS). [Figure 34] FIG. 34 shows results similar to FIG. 33 but with a larger angular separation between users. [Figure 35] 10 is a plot of SNR threshold as a function of AS for different values ​​of the user's mean angle of arrival (AOA). [Figure 36] FIG. 1 illustrates SNR thresholds for an exemplary case of five users. [Figure 37] FIG. 10 provides a comparison of SNR thresholds of BD and ASel with one and two extra antennas for two user cases. [Figure 38] FIG. 38 shows results similar to FIG. 37 but for the five user case. [Figure 39] FIG. 10 is a diagram illustrating the SNR thresholds of the BD scheme with different values ​​of AS. [Figure 40] FIG. 10 shows SNR thresholds in spatially correlated channels at AS=0.1° for BD and ASel with one and two extra antennas. [Figure 41] FIG. 10 illustrates the calculation of SNR thresholds for the two-channel scenario with AS=5°. [Figure 42] FIG. 10 illustrates the calculation of SNR thresholds for the two-channel scenario with AS=10°. [Figure 43] 10 shows the SNR threshold as a function of the number of users (M) and the angular spread (AS) for BD and ASel schemes with one and two extra antennas, respectively. [Figure 44] 10 shows the SNR threshold as a function of the number of users (M) and the angular spread (AS) for BD and ASel schemes with one and two extra antennas, respectively. [Figure 45] FIG. 1 illustrates a receiver equipped with a frequency offset estimator / corrector. [Figure 46] FIG. 1 illustrates a DIDO 2x2 system model according to one embodiment of the present invention. [Figure 47] FIG. 1 illustrates a method according to one embodiment of the present invention. [Figure 48] 10 shows the SER results of the DIDO2×2 system with and without frequency offset. [Figure 49] FIG. 10 compares the performance of different DIDO schemes with respect to SNR threshold. [Figure 50] FIG. 10 compares the amount of overhead required for different embodiments of the method. [Figure 51] FIG. 10 shows a simulation with a small frequency offset of fmax=2 Hz and no integer offset correction. [Figure 52] FIG. 10 shows the results of powering down the integer offset estimator. DETAILED DESCRIPTION OF THE INVENTION

[0023] In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced without some of these specific details. In other instances, well-known structures and devices are shown in block diagram form to avoid obscuring the underlying principles of the present invention.

[0024] 1 shows a prior art MIMO system having a transmit antenna 104 and a receive antenna 105. Such a system can typically provide up to three times the throughput that would be achievable over the available channel. There are several different approaches to implementing the details of such a MIMO system that are described in the published literature for this invention, and the following description focuses on one such approach.

[0025] Before data is transmitted in the MIMO system of FIG. 1, the channel is "characterized." This is accomplished by first transmitting a "training signal" from each of the transmit antennas 104 to each of the receivers 105. The training signals are generated by the coding and modulation subsystem 102, converted to analog by a D / A converter (not shown), and then subsequently converted from baseband to RF by each transmitter 103. Each receive antenna 105, coupled to a respective RF receiver 106, receives and converts each training signal to baseband. The baseband signals are converted to digital by a D / A converter (not shown), and the signal processing subsystem 107 characterizes the training signals. The characterization of each signal can include many factors, including, for example, phase and amplitude relative to a reference internal to the receiver, an absolute reference, a relative reference, inherent noise, or other factors. The characterization of each signal is typically defined as a vector that characterizes the phase and amplitude changes of some aspect of the signal when transmitted across the channel. For example, in a quadrature amplitude modulation (QAM) modulated signal, the characterization could be a vector of phase and amplitude offsets of several multipath images of the signal. As another example, in an Orthogonal Frequency Division Multiplexing (OFDM) modulated signal, it may be a vector of phase and amplitude offsets of some or all of the individual sub-signals within the OFDM spectrum.

[0026] The signal processing subsystem 107 stores the channel characterizations received by each receive antenna 105 and the corresponding receiver 106. After all three transmit antennas 104 have completed their training signal transmissions, the signal processing subsystem 107 stores three channel characterizations for each of the three receive antennas 105, thus resulting in a 3x3 matrix 108 designated as the channel characterization matrix "H". Each individual matrix element H i、j is the channel characterization of the training signal transmission of transmit antenna 104i received by receive antenna 105j (generally a vector as described above).

[0027] At this point, the signal processing subsystem 107 transposes the matrix H 108 to obtain H -1 and waits for the transmission of real data from the transmit antennas 104. Note that various prior art MIMO techniques described in the available literature can be utilized to ensure that the H matrix 108 is transposable.

[0028] In operation, a payload of data to be transmitted is presented to data input subsystem 100. It is then split into three parts by splitter 101 before being presented to coding and modulation subsystem 102. For example, if the payload is the ASCII bits of "abcdef," it may be split by splitter 101 into three sub-payloads of ASCII bits for "ad," "be," and "cf." Each of these sub-payloads is then individually presented to coding and modulation subsystem 102.

[0029] Each subpayload is individually coded using a coding system that is suited to both the statistical independence and error correction capabilities of each signal. These include, but are not limited to, Reed-Solomon coding, Viterbi coding, and Turbo Codes. Finally, each of the three coded subpayloads is modulated with a modulation scheme appropriate for the channel. Examples of modulation techniques include differential phase shift keying (DPSK) modulation, 64-QAM modulation, and OFDM. Note that the diversity gain provided by MIMO enables higher-order modulation constellations than would otherwise be achievable within a single-input, single-output (SISO) system using the same channel. Each coded and modulated signal is then transmitted through a unique antenna 104 after digital-to-analog conversion by a digital-to-analog conversion unit (not shown) and RF generation by each transmitter 103.

[0030] Assuming adequate spatial diversity exists between the transmit and receive antennas, each of the receive antennas 105 will receive a different combination of the three transmitted signals from antenna 104. Once received, each signal is converted to baseband by a respective RF receiver 106 and digitized by an A / D converter (not shown). n is the signal received by the nth receiving antenna 105, and X n If ∑ n ⁢ ... n ⁢ ⁢ is the signal transmitted by the nth transmit antenna 104 and N is the noise, this can be described by the following three equations: y1= x1H 11 + x2H 12 + x3H 13 + N y2= x1H 21 + x2H 22 + x3H 23 + N y3= x1H 31 + x2H 32 + x3H 33 + N

[0031] Considering that this is a system of three equations with three unknowns, it is a linear algebra problem for the signal processing subsystem 107 to derive x1, x2, and x3 (assuming N is low enough to allow decoding of the signal). x1= y1H -1 11 + y2H -1 12 + y3H -1 13 x2= y1H -1 21 + y2H -1 22 + y3H -1 23 x3= y1H -1 31 + y2H -1 32 + y3H -1 33

[0032] 3 transmit signals xn With , derived in this manner, it is demodulated, decoded, and error corrected by signal processing subsystem 107 to recover the three bit streams originally separated by splitter 101. These bit streams are combined in combiner unit 108 and output as a single data stream at data output 109. Assuming the robustness of the system is able to overcome noise impairments, data output 109 will produce the same bit stream that was introduced at data input 100.

[0033] While the prior art systems described above are typically practical for up to four antennas, and perhaps as many as ten antennas, they become impractical for large numbers of antennas (e.g., 25, 100, or 1000) for reasons explained in the "Background" section of this disclosure.

[0034] Typically, such prior art systems are bidirectional, and the return path is implemented in exactly the same way, but reversed, with each side of the communication channel having both a transmitting subsystem and a receiving subsystem.

[0035] FIG. 2 illustrates one embodiment of the present invention, where a base station (BS) 200 comprises a Wide Area Network (WAN) interface (e.g., to the Internet via a T1 or other high-speed connection) 201 and is equipped with several (N) antennas 202. For the time being, we use the term "base station" to refer to any wireless station that communicates wirelessly with a set of clients from a fixed location. Examples of base stations are access points in a wireless local area network (WLAN) or WAN antenna tower or antenna array. There are several client devices 203-207, each with a single antenna, that are served wirelessly by the base station 200. For purposes of this example, it is easiest to think of a base station located in an office environment serving client devices 203-207 that are wireless network-equipped personal computers, but this architecture applies to many applications where a base station serves wireless clients, both indoors and outdoors. For example, a base station could be based on a cell phone tower or a television broadcast tower. In one embodiment, base station 200 is positioned on the ground and configured to transmit upward at HF ​​frequencies (e.g., frequencies up to 24 MHz) to bounce signals off the ionosphere, as described in currently pending U.S. patent application Ser. No. 10 / 817,731, entitled "System and Method for Improving Near-Normal Incidence Skywave (NVIS) Communications Using Space-Time Coding," filed April 20, 2004, which is assigned to the assignee of the present application and is incorporated herein by reference.

[0036] Certain details relating to the base station 200 and client devices 203-207 described above are for illustrative purposes only and are not required to follow the underlying principles of the present invention. For example, the base station can be connected to a variety of different types of wide area networks through WAN interface 201, including application-specific wide area networks such as those used for digital video distribution. Similarly, the client devices can be any type of wireless data processing and / or communication device, including, but not limited to, mobile phones, personal digital assistants (PDAs), receivers, and wireless cameras.

[0037] In one embodiment, the base station's n antennas 202 are spatially separated so that each transmits and receives spatially uncorrelated signals, as if the base station were a prior art MIMO transceiver. As noted above in the "Background," antennas spaced within λ / 6 (i.e., 1 / 6 wavelength) have successfully achieved increased throughput from MIMO, but experiments have shown that overall, the greater the spacing between these base station antennas, the better the system performance, with λ / 2 being a desirable minimum. Of course, the underlying principles of the present invention are not limited to any particular separation between antennas.

[0038] Note that a single base station 200 will very likely have antennas spaced very far apart. For example, in the HF spectrum, antennas may be spaced 10 meters apart or more (e.g., in the NVIS example above). If 100 such antennas were used, the base station's antenna array could quite possibly span several square kilometers.

[0039] In addition to spatial diversity techniques, one embodiment of the present invention polarizes signals to increase the effective throughput of the system. Increasing channel capability through polarization is a technique used by satellite television providers for many years. Polarization can be used to place multiple (e.g., three) base station or user antennas very close to each other and still be spatially uncorrelated. While traditional RF systems typically benefit from diversity in only two dimensions of polarization (e.g., x and y), the architecture described herein can additionally benefit from diversity in three dimensions of polarization (x, y, and z).

[0040] In addition to spatial and polarization diversity, one embodiment of the present invention uses antennas with near-orthogonal radiation patterns to improve link performance through pattern diversity. Pattern diversity can improve the performance and error rate performance of MIMO systems, and its advantages over other antenna diversity techniques are shown in the following papers:

[17] L. Dong, H. Ling, and R.W. Heath, Jr., "Multiple-input multiple-output wireless communication system using antenna pattern diversity," Proceedings of the IEEE Global Telecom Conf., Vol. 1, pp. 997-1001, November 2002.

[18] R. Vaughan, "Switched Parasitic Elements for Antenna Diversity," IEEE Transactions on Antenna Propagation, Vol. 47, pp. 399-405, February 1999.

[19] P. Mattheijssen, M. H. A. Herben, G. Dolmans, and L. Leyten, "Antenna Pattern Diversity Versus Spatial Diversity for Handheld Use," IEEE Transactions on Veh. Technol., Vol. 53, pp. 1035-1042, July 2004.

[20] C.B. Dietrich Jr., K. Dietze, J.R. Nealy, and W.L. Tutzman, "Spatial, Polarization, and Pattern Diversity for Wireless Handheld Terminals," Proceedings of the IEEE Antennas and Propagation Symposium, Vol. 49, pp. 1271-1281, September 2001.

[21] A. Forenza and R.W. Heath, Jr., "Pattern diversity benefits from a two-element array of circular patch antennas in indoor clustered MIMO channels," IEEE Transactions on Communications, Vol. 54, No. 5, pp. 943-954, May 2006. Polarization can be used to place multiple base station or user antennas very close to each other, yet still be spatially uncorrelated.

[0041] Figure 3 shows further details of one embodiment of base station 200 and client devices 203-207 shown in Figure 2. For simplicity, base station 300 is shown as having only three antennas 305 and only three client devices 306-308. However, it should be noted that embodiments of the invention described herein can be implemented with a virtually unlimited number of antennas 305 (i.e., limited only by available space and noise) and client devices 306-308.

[0042] FIG. 3 is similar to the prior art MIMO architecture shown in FIG. 1 in that both systems have three antennas on each side of the communication channel. A notable difference is that in the prior art MIMO system, the three antennas 105 on the right side of FIG. 1 are all at a fixed distance from each other (e.g., integrated on a single device), and the received signals from each of the antennas 105 are processed jointly in the signal processing subsystem 107. In contrast, in FIG. 3, each of the three antennas 309 on the right side of the diagram is coupled to a different client device 306-308, each of which may be distributed anywhere within the range of the base station 305. Thus, the signal received by each client device is processed independently from the other two received signals in its coding, modulation, and signal processing subsystem 311. Thus, in contrast to a multiple-input (i.e., antenna 105), multiple-output (i.e., antenna 104) MIMO system, FIG. 3 illustrates a multiple-input (i.e., antenna 305), distributed output (i.e., antenna 305) system, hereinafter referred to as a "MIMO" system.

[0043] This application uses different terminology than the prior applications to better align with academic and industry practices. In the previously referenced currently pending U.S. patent application Ser. No. 10 / 817,731, filed April 20, 2004, entitled "System and Method for Improving Near-Normal Incidence Skywave (NVIS) Communications Using Space-Time Coding," and U.S. patent application Ser. No. 10 / 902,978, filed July 30, 2004, of which this application is a continuation-in-part, the meanings of "input" and "output" (in the context of SIMO, MISO, DIMO, and MIDO) are reversed from the usage of the terms in this application. In the prior applications, "input" referred to the radio signal as it was input to a receive antenna (e.g., antenna 309 in FIG. 3 ), and "output" referred to the radio signal as it was output by a transmit antenna (e.g., antenna 305). In academia and the wireless industry, the opposite meanings of "input" and "output" are commonly used, with "input" referring to the radio signal as it enters the channel (i.e., the transmitted radio signal from antenna 305) and "output" referring to the radio signal as it leaves the channel (i.e., the radio signal received by antenna 309). This application uses these terms inversely to the applications cited earlier in this paragraph. Therefore, equivalence of the following terms should be drawn between the applications.

[0044] (table) TIFF2025183211000002.tif30152

[0045] The MIDO architecture shown in FIG. 3 provides similar performance gains as MIMO over SISO systems for a given number of transmit antennas. However, the difference between MIMO and the specific MIDO embodiment shown in FIG. 3 is that each MIDO client device 306-308 requires only one receive antenna to provide the performance gains achieved by multiple base station antennas, whereas for MIMO, each client device requires as many receive antennas as the desired performance factor is achieved. Given the usual practical limitations on the number of antennas that can be installed on a client device (as discussed in the "Background"), this typically limits MIMO systems to between four and ten antennas (and performance factors of four to ten). Because base station 300 typically serves many client devices from a fixed, powered location, it is practical to extend the antenna count to much more than ten and separate the antennas by an appropriate distance to provide spatial diversity. As shown, each antenna is equipped with a transceiver 304 and coding, modulation, and some processing power of the signal processing unit 303. Notably, in this embodiment, regardless of the scalability of the base station 300, only one antenna is required for each client device 306-308, thus keeping the cost of each individual user client device 306-308 low and allowing the cost of the base station 300 to be shared among a large base of users.

[0046] Examples of how MIDO transmissions from base station 300 to client devices 306-308 can be accomplished are shown in FIGS.

[0047] In one embodiment of the present invention, the channel is characterized before MIMO transmission begins. As in a MIMO system, training signals are transmitted one by one from each of the antennas 405 (in the embodiment described herein). FIG. 4 shows the first training signal transmission, but there are three separate transmissions in total for the three antennas 405. Each training signal is generated by the coding, modulation, and signal processing subsystem 403, converted to analog via a D / A converter, and transmitted as RF through each RF transceiver 404. A variety of different coding, modulation, and signal processing techniques can be used, including, but not limited to, those mentioned above (e.g., Reed Solomon, Viterbi coding; QAM, DPSK, QPSK modulation, etc.).

[0048] Each client device 406-408 receives the training signal via antenna 409 and converts it to baseband via transceiver 410. An A / D converter (not shown) converts the signal to digital form, which is then processed by the respective coding, modulation, and signal processing subsystem 411. The signal characterization logic 320 then characterizes the resulting signal (e.g., by determining phase and amplitude distortions as described above) and stores the characterization in memory. This characterization process is similar to that of prior art MIMO systems, with the notable difference that each client device only calculates a characterization vector for its own antenna rather than for all n antennas. For example, the coding, modulation, and signal processing subsystem 420 of client device 406 is initialized with the known pattern of the training signal (by receiving it in a transmitted message during manufacture or through a separate initialization process). When antenna 405 transmits a training signal with this known pattern, coding, modulation, and signal processing subsystem 420 uses correlation techniques to find the strongest received pattern of the training signal, stores the phase and amplitude offsets, and then subtracts this pattern from the received signal. It then finds the second strongest received pattern that correlates to the training signal, stores the phase and amplitude offsets, and then subtracts this second strongest pattern from the received signal. This process continues until some fixed number of phase and amplitude offsets have been stored (e.g., 8), or until the number of detectable training signal patterns is below a predetermined noise level. This vector of phase / amplitude offsets is represented by element H of vector 413. 11 At the same time, the coding, modulation, and signal processing subsystems of client devices 407 and 408 perform the same processing to convert their vector elements H 21 and H 31 Generate.

[0049] The memory in which the characterization is stored can be non-volatile memory such as flash memory or volatile memory such as a hard drive and / or random access memory (e.g., SDRAM, RDAM). Furthermore, different client devices may simultaneously use different types of memory to store characterization information (e.g., a PDA may use flash memory, while a laptop personal computer may use a hard drive). The underlying principles of the present invention are not limited to any particular type of storage mechanism on the various client devices or base stations.

[0050] As mentioned above, depending on the scheme used, each client device 406-408 has only one antenna, and therefore each stores a 1x3 row 413-415 of the H matrix. FIG. 4 illustrates the stage after the first training signal transmission, where the first column of the 1x3 rows 413-415 stores channel characterization information for the first of the three base station antennas 405. The remaining two columns are stored after channel characterization of the next two training signal transmissions from the remaining two base station antennas. Note that for illustrative purposes, the three training signals are transmitted at separate times. If the three training signal patterns are selected to be uncorrelated with each other, they can be transmitted simultaneously, thus reducing training time.

[0051] As shown in Figure 5, after all three pilot transmissions are complete, each client device 506-508 transmits a stored 1x3 row 513-515 of matrix H to base station 500. For simplicity, Figure 5 shows only one client device 506 transmitting characterization information. An appropriate modulation scheme for the channel (e.g., DPSK, 64QAM, OFDM) combined with appropriate error correction coding (e.g., Reed Solomon, Viterbi, and / or Turbo codes) can be used to ensure that base station 500 receives the data accurately in rows 513-515.

[0052] 5, it is sufficient for one antenna and transceiver at base station 500 to receive each 1x3 row 513-515 transmission. However, utilizing many or all of the antennas 505 and transceivers 504 to receive each transmission (i.e., utilizing conventional single-input multiple-output (SIMO) processing techniques within coding, modulation, and signal processing subsystem 503) can provide a better signal-to-noise ratio (SNR) than utilizing a single antenna 505 and transceiver 504 under certain conditions.

[0053] The coding, modulation, and signal processing subsystem 503 of the base station 500 receives 1x3 rows 513-515 from each client device 507-508 and stores them in a 3x3 H matrix 516. As with the client devices, the base station can store the matrix 516 using a variety of different storage technologies, including, but not limited to, non-volatile mass storage memory (e.g., a hard drive) and / or volatile memory (e.g., SDRAM). Figure 5 shows the stage at which the base station 500 has received and stored 1x3 row 513 from client device 509. 1x3 rows 514 and 515 can be transmitted and stored in H matrix 516 as they are received from the remaining client devices until the entire H matrix 516 is stored.

[0054] One embodiment of a MIDO transmission from a base station 600 to client devices 606-608 will now be described with reference to FIG. 6. Because each client device 606-608 is an independent device, typically each device receives a different data transmission. Accordingly, one embodiment of a base station 600 includes a router 602 communicatively positioned between a WAN interface 601 and a coding, modulation, and signal processing subsystem 603, which provides multiple data streams (formatted into bit streams) from the WAN interface 601, sent as separate bit streams u1-u3 intended for each client device 606-608, respectively. Various known routing techniques can be used by the router 602 for this purpose.

[0055] The three bit streams u1-u3 shown in FIG. 6 are then sent to the coding, modulation, and signal processing subsystem 603, where they are coded into statistically distinct error-correcting streams (e.g., with Reed Solomon, Viterbi, or Turbo Codes) and modulated for the channel with an appropriate modulation scheme (e.g., DPSK, 64QAM, or OFDM). Additionally, the embodiment shown in FIG. 6 includes signal precoding logic 630 that uniquely codes the signal transmitted from each antenna 605 based on a signal characterization matrix 616. More specifically, rather than sending each of the three coded and modulated bit streams to a separate antenna (as is done in FIG. 1), in one embodiment, the precoding logic 630 multiplies the three bit streams u1-u3 by the inverse of the H matrix 616 to generate three new bit streams u′1-u′3. The three precoded bit streams are then converted to analog by a D / A converter (not shown) and transmitted as RF by the transceiver 604 and antenna 605.

[0056] Before describing how the bitstreams are received by the client devices 606-608, a description will be given of the operations performed by the precoding module 630. As with the previous MIMO example of FIG. 1, the coded and modulated signals for each of the three source bitstreams are n In the embodiment shown in FIG. i contains data for one of three bitstreams routed by router 602, each such bitstream intended for one of three client devices 606-608.

[0057] But each x i Unlike the MIMO example of FIG. 1, where each u is transmitted by each antenna 104, in the embodiment of the invention shown in FIG. i is received at each client device antenna 609 (plus any noise N present in the channel). To achieve this result, three antennas 605 (each of which receives vi Each output of i and a function of the H matrix that characterizes the channel of each client device. i is calculated by the precoding logic 630 in the coding, modulation, and signal processing subsystem 603 by implementing the following formula: v1= u1H -1 11 + u2H -1 12 + u3H -1 13 v2= u1H -1 21 + u2H -1 22 + u3H -1 23 v3= u1H -1 31 + u2H -1 32 + u3H -1 33

[0058] That is, after the signal is transformed by the channel, each x i Unlike MIMO, where v is calculated at the receiver, the embodiments of the invention described herein calculate each v at the transmitter before the signal is transformed by the channel. i Each antenna 609 is designed to have other u n-1 separated from the bitstream i Each transceiver 610 converts each received signal to baseband, each received signal is digitized by an A / D converter (shown here), and each coding, modulation, and signal processing subsystem 611 generates the x intended for each received signal. i It demodulates and decodes the bitstream and sends it to a data interface 612 for use by the client device (eg, by an application on the client device).

[0059] The embodiments of the present invention described herein can be implemented using a variety of different coding and modulation schemes. For example, in OFDM implementations where the frequency spectrum is separated into multiple subbands, the techniques described herein can be used to characterize each individual subband. However, as noted above, the underlying principles of the present invention are not limited to any particular modulation scheme.

[0060] If the client devices 606-608 are portable data processing devices such as PDAs, laptop personal computers, and / or wireless telephones, the channel characterization may change frequently as the client devices may move from one location to another. Therefore, in one embodiment of the present invention, the channel characterization matrix 616 at the base station is continuously updated. In one embodiment, the base station 600 periodically (e.g., every 250 milliseconds) sends a new training signal to each client device, and each client device continually transmits its channel characterization vector to the base station 600 to ensure that the channel characterization remains accurate (e.g., as the environment changes to affect the channel or the client device moves). In one embodiment, the training signal is interleaved within the actual data signal sent to each client device. Typically, the training signal has a much lower throughput than the data signal, and therefore has little impact on the overall throughput of the system. Therefore, in this embodiment, the channel characterization matrix 616 can be continuously updated as the base station actively communicates with each client device, thereby maintaining accurate channel characterization even as the client device moves from one location to another or as the environment changes to affect the channel.

[0061] One embodiment of the present invention, shown in Figure 7, uses MIMO techniques to improve the upstream communication channel (i.e., the channel from client devices 706-708 to base station 700). In this embodiment, the channel from each of the client devices is continuously analyzed and characterized by upstream channel characterization logic 741 within the base station. More specifically, each of the client devices 706-708 transmits a training signal to the base station 700 (e.g., as in a typical MIMO system) which the channel characterization logic 741 analyzes to generate an NxM channel characterization matrix 741, where N is the number of client devices and M is the number of antennas used by the base station. The embodiment shown in Figure 7 uses three antennas 705 at the base station and three client devices 706-708, and therefore a 3x3 channel characterization matrix 741 is stored at the base station 700. The MIMO upstream transmission shown in FIG. 7 can be used by client devices to transmit data to the base station 700 and to transmit channel characterization vectors to the base station 700 as shown in FIG. 5, but unlike the embodiment shown in FIG. 5 in which the channel characterization vectors for each client device are transmitted at separate times, the method shown in FIG. 7 allows for simultaneous transmission of channel characterization vectors from multiple client devices to the base station 700, thus drastically reducing the impact of the channel characterization vectors on the return channel throughput.

[0062] As mentioned above, the characterization of each signal can include many factors, including, for example, phase and amplitude relative to a reference internal to the receiver, an absolute reference, a relative reference, inherent noise, or other factors. For example, in a quadrature amplitude modulation (QAM) modulated signal, the characterization can be a vector of phase and amplitude offsets of several multipath images of the signal. As another example, in an orthogonal frequency division multiplexing (OFDM) modulated signal, it can be a vector of phase and amplitude offsets of some or all of the individual sub-signals within the OFDM spectrum. The training signal can be generated by each client device's encoding and modulation subsystem 711, converted to analog by a D / A converter (not shown), and then converted from baseband to RF by each client device's transmitter 709. In one embodiment, to ensure that the training signal is synchronized, the client device transmits the training signal only when requested by the base station (e.g., in a round-robin manner). Furthermore, the training signal can be interleaved within or transmitted simultaneously with the actual data signal sent by each client device. Thus, even if the client devices 706-708 are mobile, training signals can be continuously transmitted and analyzed by the upstream channel characterization logic 741, thus ensuring that the channel characterization matrix 741 remains up to date.

[0063] The total channel capability supported by the above-described embodiments of the present invention can be defined as min(N,M), where M is the number of client devices and N is the number of base station antennas. That is, the capability is limited by the number of antennas on the base station or client side. Therefore, one embodiment of the present invention uses synchronization techniques to ensure that only min(N,M) antennas are transmitting and receiving at any given time.

[0064] In a typical scenario, the number of antennas 705 on a base station 700 is less than the number of client devices 706-708. An exemplary scenario is illustrated in FIG. 8, which shows five client devices 804-808 communicating with a base station having three antennas 802. In this embodiment, after determining the total number of client devices 804-808 and retrieving necessary channel characterization information (e.g., as described above), the base station 800 selects a first group of three clients 810 to communicate with (three clients in the example because min(N, M)=3). After communicating with the first group of clients 810 for a specified period of time, the base station then selects another group of three clients 811 to communicate with. To evenly distribute the communication channel, the base station 800 selects two client devices 807, 808 that were not included in the first group. Additionally, because an extra antenna is available, the base station 800 selects yet another client device 806 that was included in the first group. In one embodiment, base station 800 cycles through the groups of clients in this manner so that each client is allocated substantially the same amount of throughput over time. For example, to allocate throughput evenly, the base station may then select any combination of three client devices except for client device 806 (i.e., because client device 806 was engaged in communication with the base station through the first two cycles).

[0065] In one embodiment, in addition to standard data communications, the base station can transmit training signals to each client device using the techniques described above, and receive training signals and signal characterization data from each client device.

[0066] In one embodiment, certain client devices or groups of client devices may be assigned different levels of throughput. For example, client devices may be prioritized to ensure more communication cycles (i.e., more throughput) for relatively higher priority client devices than for relatively lower priority client devices. The "priority" of a client device may be selected based on several variables, including, for example, a specified level of user subscription to wireless service (e.g., a user may be willing to pay for additional throughput) and / or the type of data being communicated to and from the client device (e.g., real-time communications such as telephone voice and video may be prioritized over non-real-time communications such as email).

[0067] In one embodiment, the base station dynamically allocates throughput based on the current load required by each client device. For example, if client device 804 is streaming live video while other devices 805-808 are performing non-real-time functions such as email, base station 800 may allocate relatively more throughput to this client 804. However, it should be noted that the underlying principles of the present invention are not limited to any particular throughput technique.

[0068] As shown in FIG. 9, two client devices 907, 908 can be located close enough that the client channel characterizations are virtually identical. Therefore, the base station receives and stores virtually identical channel characterization vectors for the two client devices 907, 908, and thus cannot create unique spatially dispersed signals for each client device. Therefore, in one embodiment, the base station will ensure that any two or more client devices that are near each other are assigned to different groups. In FIG. 9, for example, base station 900 first communicates with a first group 910 of client devices 904, 905, and 908, and then with a second group 911 of client devices 905, 906, and 907, ensuring that client devices 907 and 908 are in different groups.

[0069] Alternatively, in one embodiment, base station 900 communicates with client devices 907 and 908 simultaneously, but multiplexes the communication channels using known channel multiplexing methods. For example, the base station may split a single spatially correlated signal between client devices 907 and 908 using time division multiplexing (TDM), frequency division multiplexing (FDM), or code division multiple access (CDMA) techniques.

[0070] Although each of the client devices described above is equipped with one antenna, the underlying principle of the present invention of using client devices with multiple antennas to increase throughput can be used. For example, when used on the wireless system described above, a client with two antennas will provide a two-fold increase in throughput. A client with three antennas will provide a three-fold increase in throughput, and so on (i.e., assuming sufficient spatial and angular separation between the antennas). A base station can apply the same general rules when cycling through client devices with multiple antennas. For example, it can treat each antenna as a separate client and allocate throughput to that "client" as it would any other client (e.g., ensuring that each client is given appropriate or equal communication time).

[0071] As mentioned above, one embodiment of the present invention increases the signal-to-noise ratio and throughput in a near-normal incidence skywave (NVIS) system using MIDO and / or the above-mentioned MIMO signal transmission techniques. Referring to FIG. 10 , in one embodiment of the present invention, a first NVIS station 1001 equipped with a matrix of N antennas 1002 is configured to communicate with M client devices 1004. The NVIS antenna 1002 and the antennas of the various client devices 1004 transmit signals upward to approximately 15 degrees vertical to provide the desired NVIS and minimize ground wave interference effects. In one embodiment, the antenna 1002 and the client devices 1004 support multiple independent data streams 1006 using various MIDO and above-mentioned MIMO techniques at designated frequencies within the NVIS spectrum (e.g., at the carrier frequency or below 23 MHz, but generally below 10 MHz), thus significantly increasing throughput at the designated frequencies (i.e., by a factor proportional to the number of statistically independent data streams).

[0072] NVIS antennas serving a given station may be physically separated from one another by great distances. Given the long wavelengths (<10 MHz) and the long distances the signals travel (as long as 300 miles round trip), physical separation of antennas on the order of hundreds of yards and even miles can provide diversity benefits. In such situations, individual antenna signals can be returned to a centralized location for processing using conventional wired or wireless communication systems. Alternatively, each antenna may have local functionality to process its signal and then return the data to a centralized location using conventional wired or wireless communication systems. In one embodiment of the present invention, the NVIS station 1001 has a broadband link 1015 to the Internet 1010 (or other wide area network), thereby providing remote high-speed wireless network access to client devices 1003.

[0073] In one embodiment, the base station and / or users can utilize the polarization / pattern diversity techniques described above to reduce array size and / or user distance while providing diversity and increased throughput. As an example, in a MIDO system with HF transmissions, users can be in the same location, yet the signals can be uncorrelated due to polarization / pattern diversity. In particular, by using pattern diversity, one user can be communicating to a base station over terrestrial radio, while another user can be communicating through NVIS.

[0074] Additional Embodiments of the Invention 1. DIDO-OFDM Precoding with I / Q Imbalance One embodiment of the present invention employs a system and method for correcting in-phase and quadrature (I / Q) imbalance in a distributed input distributed output (DIDO) system using orthogonal frequency division multiplexing (OFDM). Briefly, according to this embodiment, a user device estimates a channel and feeds this information back to a base station. The base station calculates a precoding matrix to cancel inter-carrier and inter-user interference caused by the I / Q imbalance. Parallel data streams are transmitted to multiple user devices through DIDO precoding, and the user devices demodulate the data through zero-forcing (ZF), minimum mean square error (MMSE), or maximum likelihood (ML) receivers to suppress the residual interference.

[0075] As will be described in detail below, some of the significant features of this embodiment of the present invention include, but are not limited to:

[0076] Precoding to cancel inter-carrier interference (ICI) (due to I / Q mismatch) caused by mirror tones in OFDM systems.

[0077] Precoding to cancel inter-user interference and ICI (due to I / Q mismatch) in DIDO-OFDM systems.

[0078] A technique to cancel ICI (due to I / Q mismatch) through a ZF receiver in DIDO-OFDM systems using a block diagonalization (BD) precoder.

[0079] A technique to cancel inter-user interference and ICI (due to I / Q mismatch) in DIDO-OFDM systems through precoding (at the transmitter) and ZF or MMSE filters (at the receiver).

[0080] A technique to cancel inter-user interference and ICI (due to I / Q mismatch) through precoding (at the transmitter) and non-linear detectors such as maximum likelihood (ML) detectors (at the receiver) in DIDO-OFDM systems.

[0081] Using precoding based on channel state information to cancel inter-carrier interference (ICI) due to mirror tones (due to I / Q mismatch) in OFDM systems.

[0082] Using precoding based on channel state information to cancel inter-carrier interference (ICI) due to mirror tones (due to I / Q mismatch) in DIDO-OFDM systems.

[0083] Use of an I / Q mismatch-aware DIDO precoder at the base station and an IQ-aware DIDO receiver at the user terminal.

[0084] Use of an I / Q mismatch-aware DIDO precoder at the base station, an I / Q-aware DIDO receiver at the user terminal, and an I / Q-aware channel estimator.

[0085] Use of an I / Q mismatch-aware DIDO precoder at the base station, an I / Q-aware DIDO receiver and I / Q-aware channel estimator at the user terminal, and an I / Q-aware DIDO feedback generator to send channel state information from the user terminal to the base station.

[0086] Use of an I / Q mismatch-aware DIDO precoder at the base station and an I / Q-aware DIDO configurer that uses I / Q path information to perform functions including user selection, adaptive coding and modulation, space-time frequency mapping, or precoder selection.

[0087] Use of an I / Q-aware DIDO receiver to cancel ICI (due to I / Q mismatch) through a ZF receiver in a DIDO-OFDM system using a block diagonalization (BD) precoder.

[0088] The use of an I / Q-aware DIDO receiver to cancel ICI (due to I / Q mismatch) through precoding (at the transmitter) and a nonlinear detector such as a maximum likelihood (ML) detector (at the receiver) in a DIDO-OFDM system.

[0089] Use of I / Q aware DIDO receivers to cancel ICI (due to I / Q mismatch) through ZF or MMSE filters in DIDO-OFDM systems.

[0090] a. background The transmit and receive signals in a typical wireless communication system consist of in-phase and quadrature (I / Q) components. In practical systems, the in-phase and quadrature components can be distorted due to imperfections in mixing and baseband operation. These distortions manifest as I / Q phase, gain, and delay mismatch. Phase imbalance is caused by sines and cosines in the modulator / demodulator not being perfectly quadrature. Gain imbalance is caused by different amplifications between the in-phase and quadrature components. An additional distortion, delay imbalance, can exist due to differences in delay between the I and Q rails in analog circuits.

[0091] In Orthogonal Frequency Division Multiplexing (OFDM), I / Q imbalance causes inter-carrier interference (ICI) from mirror tones. This effect has been studied in the literature, and methods for correcting I / Q mismatch in single-input, single-output SISO-OFDM systems are discussed in M.D. Benedetto and P. Mandarini, "Analysis of the Impact of I / Q Baseband Filter Mismatch in OFDM Modems," Wireless Personal Communications, pp. 175-186, 2000; S. Schuchert and R. Hasholzner, "A Novel I / Q Imbalance Correction Scheme for Receiving OFDM Signals," IEEE Transactions on Consumer Electronics, August 2001; M. Valkama, M. Renfors, and V. Koivunen. "Advanced Methods for I / Q Imbalance Compensation in Communication Receivers," IEEE Transactions on Signal Processing, October 2001; R. Rao and B. Daneshrad, "Analysis of I / Q Mismatch and Cancellation Schemes for OFDM Systems," IST Mobile Communications Summit, June 2004; A. Tarighat, R. Bagheri, and A. H. Sayed, "Compensation Schemes and Performance Analysis of IQ Imbalance in OFDM Receivers," IEEE Transactions on Signal Processing [see also IEEE Transactions on Acoustics, Speech, and Signal Processing], Vol. 53, pp. 3257-3268, August 2005.

[0092] An extension of this work to multiple-input multiple-output MIMO-OFDM systems, for spatial multiplexing (SM), is given in R. Rao and B. Daneshrad, "I / Q Mismatch Cancellation in MIMO-OFDM Systems," in Personal, Indoor, and Mobile Radio Communications, 2004, PIMRC2004, 15th IEEE International Symposium, Vol. 4, 2004, pp. 2710-2714; R. M. Rao, W. Zhu, S. Lang, C. Oberli, D. Browne, J. Bhatia, J. Frigon, J. Wang, P. Gupta, H. Lee, D. N. Liu, S. G. Wong, M. Fitz, B. Daneshrad, and O. Takeshita, "Wireless Communications Research Teachings," in "A Multi-Antenna Testbed for Software-Defined Radio OFDM Communication," IEEE Communications Journal, Vol. 42, No. 12, pp. 72-81, December 2004; S. Lang, M.R. Rao, and B. Daneshrad, "Design and Development of a 5.25 GHz Software-Defined Radio OFDM Communication Platform," IEEE Communications Journal, Vol. 42, No. 6, pp. 6-12, June 2004; and for Orthogonal Space-Time Block Codes (OSTBC), A. Tarighat and A.H. Sayed, "MIMO-OFDM Receiver for Systems with IQ Imbalance," IEEE Transactions on Signal Processing, Vol. 53, pp. 3583-3596, September 2005.

[0093] Unfortunately, there is currently no literature on how to correct I / Q gain and phase imbalance errors in distributed input distributed output (DIDO) communication systems. The embodiments of the present invention described below provide solutions to these problems.

[0094] A DIDO system consists of a single base station with distributed antennas that transmit parallel data streams (through precoding) to multiple users to improve downlink throughput while utilizing the same radio resources (i.e., the same slot duration and frequency band) as a traditional SISO system. A detailed description of a DIDO system is provided in U.S. patent application Ser. No. 10 / 902,978, filed July 30, 2004 (previous application), entitled "System and Method for Distributed Input Distributed Output Wireless Communications," by S.G. Perlman and T. Cotter, which is assigned to the assignee of the present application and incorporated herein by reference.

[0095] There are many ways to implement a DIDO precoder. One solution is described in QH Spencer, ALS Windlehurst, and M. Haardt, "Zero-Forcing Methods for Downlink Spatial Multiplexing in Multiuser MIMO Channels," IEEE Transactions on Signal Processing, Vol. 52, pp. 461-471, February 2004; K.K. Wong, R.D. Murch, and K.B. Letaief, "Joint Channel Diagonalization for Multiuser MIMO Antenna Systems," IEEE Transactions on Wireless Communications, Vol. 2, pp. 773-786, July 2003; L.U. Choi and R.D. Murch, "Transmit Preprocessing Techniques for Multiuser MIMO Systems Using Decomposition Techniques," IEEE Transactions on Wireless Communications, Vol. 3, pp. 20-24, January 2004; Z. Shen, J.G. Andrews, and R.W. He Heath, R., and B. L. vans, "A Low-Complexity User Selection Algorithm for Multiuser MIMO Systems with Block Diagonalization," IEEE Transactions on Signal Processing, September 2005, accepted for publication; Z. Shen, R. Chen, J. G. Andrews, R. W. Heath, and B. L. vans, "Sum Function of Multiuser MIMO Broadcast Channels with Block Diagonalization," submitted to IEEE Transactions on Wireless Communications, October 2005; and R. Chen, R. W. Heath, and J. G. Andrews, "Transmit Selection Diversity for Single-Precoded Multiuser Spatial Multiplexing Systems with Linear Receivers," accepted for publication to IEEE Transactions on Signal Processing, 2005. The I / Q compensation method presented here assumes a BD precoder but can be extended to any form of DIDO precoder.

[0096] In DIDO-OFDM systems, I / Q mismatch causes two effects: ICI and inter-user interference. The former is due to mirror tone interference, similar to SISO-OFDM systems. The latter is due to I / Q mismatch destroying the orthogonality of the DIDO precoder, resulting in inter-user interference. Both of these forms of interference can be canceled at the transmitter and receiver through the methods described herein. Three methods of I / Q compensation in DIDO-OFDM systems are described, and performance is compared with systems with and without I / Q mismatch. Results are presented based on simulations and practical measurements performed on a DIDO-OFDM prototype.

[0097] The embodiments of the present invention are extensions of the prior application. In particular, these embodiments address the following features of the prior application:

[0098] Systems such as those described in the prior application in which the I / Q rails suffer from gain and phase imbalance.

[0099] The training signal used for channel estimation is used to compute a DIDO precoder with I / Q correction at the transmitter.

[0100] The signal characterization data addresses I / Q imbalance and is used in the transmitter to compute the DIDO precoder according to the method proposed herein.

[0101] b. Embodiments of the present invention

[0102] First, the mathematical model and framework of the present invention will be described.

[0103] Before presenting the solution, it is useful to explain the central mathematical concept. We will explain this concept assuming I / Q gain and phase imbalance (phase delay is not included in the explanation and is handled automatically in the DIDO-OFDM version of the algorithm). To illustrate the basic idea, we consider two complex numbers, s = si + jS Qand h=hi+jh Q Multiply by x=h * Let s be the in-phase and quadrature components. We use subscripts to denote the in-phase and quadrature components. Recall that x I = s I h I -s Q h Q and x Q = s I h Q + s Q h I .

[0104] In matrix form, this can be rewritten as:

[0105]

number

[0106] Note the unitary transformation by the channel matrix (H). Let s be the transmitted symbols and h be the channel. The presence of I / Q gain and phase imbalance can be modeled by creating a non-unitary transformation as follows:

[0107]

number

[0108] The trick is to realize that you can write:

[0109]

number

[0110] Now, if we rewrite (A), we get

[0111]

number

[0112] It is defined as follows:

[0113]

number

[0114] and

[0115]

number

[0116] Both of these matrices have a unitary structure and can therefore be equivalently represented by a complex scalar as follows:

[0117]

number

[0118] and

[0119]

number

[0120] Using all of these observations, we can calculate the two channels, i.e., the equivalent channel h e and conjugate channel h c We can convert the valid equation back to scalar form in terms of Then the valid transformation of (5) becomes

[0121]

number

[0122] The first channel is called the equivalent channel and the second channel is called the conjugate channel. The equivalent channel is what would be observed if there were no I / Q gain and phase imbalance.

[0123] By similar reasoning, it can be shown that the input-output relations for a discrete-time MIMO NxM system with I / Q gain and phase imbalance are as follows (constructing their matrix counterparts using scalar equivalents):

[0124]

number

[0125] where t is the discrete time index:

[0126]

number

[0127] and L is the number of channel taps.

[0128] In a DIDO-OFDM system, the received signal is represented in the frequency domain. Recall from the signals and systems that:

[0129]

number

[0130] In this case,

[0131]

number

[0132] and for OFDM, the equivalent input-output relationship of the MIMO-OFDM system for subcarrier k is:

[0133]

number

[0134] where k=0, 1..., K-1 are OFDM subcarrier indices, and H e and Hc denote the matrices for the equivalent channel and the conjugate channel, respectively, and are defined as follows:

[0135]

number

[0136] and

[0137]

number

[0138] The second contribution in (1) is interference from mirror tones, which can be addressed by constructing the following stacked matrix system (carefully paying attention to the conjugates):

[0139]

number

[0140] where:

[0141]

number

[0142] and

[0143]

number

[0144] are vectors of transmitted and received symbols in the frequency domain, respectively.

[0145] This approach is used to construct the effective matrix used for DIDO operation. For example, using the DIDO 2x2 input-output relationship (assuming each user has one receive antenna), the first user device sees (in the absence of noise) the following:

[0146]

number

[0147] Meanwhile, a second user observes the following:

[0148]

number

[0149] where:

[0150]

number

[0151] are the matrices H e and H c Denote the m-th row of W∈C 4x4 is the DIDO precoding matrix. From (2) and (3), the received symbols of user m are:

[0152]

number

[0153] are two sources of interference caused by I / Q imbalance: inter-carrier interference from mirror tones (i.e., p ≠ m,

[0154]

number

[0155] ) and inter-user interference (i.e.,

[0156]

number

[0157] and

[0158]

number

[0159] ) is observed to be affected by the interference. The DIDO precoding matrix W in (3) is designed to cancel these two interference conditions.

[0160] There are several different embodiments of the DIDO precoder that can be used here, based on joint detection applied at the receiver. In one embodiment, the composite channel:

[0161]

number

[0162] (H e (m), rather than ), is used (see, e.g., QH Spencer, ALS Windlehurst, and M. Haardt, "Zero-Forcing Methods for Downlink Spatial Multiplexing in Multiuser MIMO Channels," IEEE Transactions on Signal Processing, Vol. 52, pp. 461-471, February 2004; K.K. Wong, R.D. Murch, and K.B. Letaief, "Joint Channel Diagonalization for Multiuser MIMO Antenna Systems," IEEE Transactions on Wireless Communications, Vol. 2, pp. 773-786, July 2003; L.U. Choi and R.D. Murch, "Multiuser MIMO Using Decomposition Techniques," IEEE Transactions on Wireless Communications, Vol. 1, pp. 773-786, July 2003). (See also "Transmit Preprocessing Techniques for MIMO Systems," IEEE Transactions on Wireless Communications, Vol. 3, pp. 20-24, January 2004; "Low-Complexity User Selection Algorithms for Multiuser MIMO Systems with Block Diagonalization," IEEE Transactions on Signal Processing, September 2005, accepted for publication; and "Sum Function of Multiuser MIMO Broadcast Channels with Block Diagonalization," submitted to IEEE Transactions on Wireless Communications, October 2005.) Therefore, current DIDO systems select a precoder as follows:

[0163]

number

[0164] where α i、j is a constant, and

[0165]

number

[0166] This method is useful because using this precoder it is possible to keep other aspects of the DIDO precoder the same as before, since the effects of I / Q gain and phase imbalance are completely cancelled out at the transmitter.

[0167] It is also possible to design a DIDO precoder that pre-cancels inter-user interference without pre-cancelling ICI due to IQ imbalance. With this approach, the receiver (rather than the transmitter) corrects the IQ imbalance by using one of the receive filters described below. Then, the precoding design criterion in (4) can be modified as follows:

[0168]

number

[0169]

number

[0170] and

[0171]

number

[0172] Here, for the m-th transmitted symbol,

[0173]

number

[0174] and

[0175]

number

[0176] is the received symbol vector of user m.

[0177] At the receiver, the transmitted symbol vector

number

[0178]

number

[0179] Although the ZF filter is the easiest to understand, the receiver can apply any number of other filters known to those skilled in the art. One common choice is the MMSE filter, as follows:

[0180]

number

[0181] and p is the signal-to-noise ratio. Alternatively, the receiver can perform maximum likelihood symbol detection (or a sphere decoder or an iterative variant). For example, the first user can use an ML receiver to solve the following optimization:

[0182]

number

[0183] where S is the set of all possible vectors s and depends on the constellation size. The ML receiver provides better performance at the cost of increased receiver complexity. A similar set of equations applies to the second user.

[0184] H in (6) and (7) W (1,2) and H W (2,1)Note that H is assumed to have zero input. This assumption is valid only if the transmit precoder can perfectly cancel the inter-user interference, as is the case as a criterion in (4). Similarly, H W (1,1) and H W (2,2) is diagonal if and only if the transmit precoder can perfectly cancel inter-carrier interference (ie, due to mirror tones).

[0185] 13 illustrates one embodiment of a framework for a DIDO-OFDM system with I / Q correction, including an IQ-DIDO precoder 1302 at the base station (BS), a transmission channel 1304, channel estimation logic 1306 in the user equipment, and a ZF, MMSE, or ML receiver 1308. The channel estimation logic 1306 estimates the channel H through training symbols. e (m) and H c (m) and feeds these estimates back to the precoder 1302 in the AP. The BS calculates the DIDO precoder weights (matrix W) to pre-cancel the interference due to I / Q gain and phase imbalance, as well as inter-user interference, and transmits data to the users over the wireless channel 1304. User device m uses a ZF, MMSE, or ML receiver 1308 to cancel the residual interference and demodulate the data by utilizing the channel estimates provided by unit 1304.

[0186] The following three embodiments can be used to implement this I / Q correction algorithm.

[0187] Method 1 - TX Correction: In this embodiment, the transmitter calculates the precoding matrix according to the criteria in (4). At the receiver, the user equipment uses a "simplified" ZF receiver, where H W (1,1) and H W (2,2) Assume that is a diagonal matrix. Therefore, equation (8) simplifies to:

[0188]

number

[0189] Method 2 - RX Correction: In this embodiment, the transmitter calculates the precoding matrix based on the traditional BD method described in R. Chen, R.W. Heath, and J.G. Andrews, "Transmit Selection Diversity for Single-Precoding Multiuser Spatial Multiplexing Systems with Linear Receivers," accepted for IEEE Transactions on Signal Processing, 2005, without fully canceling inter-carrier and inter-user interference, as in the reference case in (4). For this method, the precoding matrices in (2) and (3) are simplified as follows:

[0190]

number

[0191] At the receiver, the user equipment uses a ZF filter as in (8). Note that this method, unlike Method 1 above, does not pre-cancel interference at the transmitter. Therefore, it cancels inter-carrier interference at the receiver, but cannot cancel inter-user interference. Furthermore, in Method 2, the user equipment uses a ZF filter as in (8). e (m) and H c (m) In contrast to Method 1, which requires feedback of both the transmitter vector H e (m) Therefore, Method 2 is particularly suitable for DIDO systems with slow feedback channels. However, Method 2 requires slightly higher computational complexity at the user equipment to calculate the ZF receiver in (8) rather than (11).

[0192] Method 3 - TX-RX Correction: In one embodiment, the two methods above are combined. The transmitter calculates the precoding matrix as in (4), and estimates the transmitted symbols according to (8) in the receiver, and the receiver estimates the transmitted symbols according to (8).

[0193] I / Q imbalance, whether phase imbalance, gain imbalance, or delay imbalance, causes detrimental degradation of signal quality in wireless communication systems. For this reason, conventional circuit hardware has been designed to have very low imbalance. However, as discussed above, digital signal processing, in the form of transmit precoding and / or special receivers, can be used to correct this problem. One embodiment of the present invention includes a system with several new functional units, each of which is important for implementing I / Q correction in an OFDM or DIDO-OFDM communication system.

[0194] One embodiment of the present invention uses precoding based on channel state information to cancel inter-carrier interference (ICI) due to mirror tones (due to I / Q mismatch) in OFDM systems. As shown in Figure 11, the DIDO transmitter according to this embodiment includes a user selector unit 1102, multiple coding and modulation units 1104, corresponding multiple mapping units 1106, a DIDO-IQ-aware precoding unit 1108, multiple RF transmitter units 1114, a user feedback unit 1112, and a DIDO configuration unit 1110.

[0195] The user selector unit 1102 selects data associated with multiple users U1-UM based on feedback information obtained by a feedback unit 1112 and provides this information to each of multiple coding and modulation units 1104. Each coding and modulation unit 1104 codes and modulates the user's information bits and sends them to a mapping unit 1106. The mapping unit 1106 maps the input bits to complex symbols and sends the results to a DIDO-IQ-aware precoding unit 1108. The DIDO-IQ-aware precoding unit 1108 calculates DIDO-IQ-aware precoding weights using channel state information obtained from the users by the feedback unit 1112 to precode the input symbols obtained from the mapping unit 1106. Each precoded data stream is sent by the DIDO-IQ-aware precoding unit 1108 to an OFDM unit 1115, which calculates an IFFT and adds a cyclic prefix. This information is sent to a D / A unit 1116, which operates a D / A conversion and sends it to an RF unit 1114. The RF unit 1114 upconverts the baseband signal to intermediate / high frequency and sends it to a transmit antenna.

[0196] The precoder operates on regular and mirror tones, both with the goal of correcting I / Q imbalance. Any number of precoder design criteria can be used, including ZF, MMSE, or weighted MMSE designs. In a preferred embodiment, the precoder completely cancels ICI due to I / Q mismatch, so the receiver does not need to provide additional correction.

[0197] In one embodiment, the precoder uses a block diagonalization criterion to fully cancel inter-user interference, but does not fully cancel I / Q effects for each user, requiring additional receiver processing. In another embodiment, the precoder uses a zero-forcing criterion to fully cancel inter-user interference and ICI due to I / Q imbalance. This embodiment can use a conventional DIDO-OFDM processor at the receiver.

[0198] One embodiment of the present invention uses channel state information-based precoding to cancel inter-carrier interference (ICI) due to mirror tones (due to I / Q mismatch) in a DIDO-OFDM system, where each user employs an IQ-aware DIDO receiver. As shown in Figure 12, in one embodiment of the present invention, a system including a receiver 1202 includes multiple RF units 1208, corresponding multiple A / D units 1210, an IQ-aware channel estimator 1204, and a DIDO feedback generator 1206.

[0199] The RF unit 1208 receives the signal transmitted from the DIDO transmitter unit 1114, downconverts the signal to baseband, and provides the downconverted signal to the A / D unit 1210. The A / D unit 1210 then converts the signal from analog to digital and sends it to the OFDM unit 1213. The OFDM unit 1213 performs an FFT to remove the cyclic prefix and transfer the signal to the frequency domain. During the training period, the OFDM unit 1213 sends its output to the IQ-aware channel estimator 1204, which calculates channel estimates in the frequency domain. Alternatively, the channel estimates can be calculated in the time domain. During the data period, the OFDM unit 1213 sends its output to the IQ-aware receiver unit 1202, which demodulates / decodes the signal to calculate an IQ receiver and obtain data 1214. The IQ-aware channel estimator 1204 sends the channel estimates to a DIDO feedback generator 1206 which can quantize the channel estimates and send them back to the transmitter over the feedback control channel 1112 .

[0200] The receiver 1202 shown in Figure 12 can operate based on any number of criteria known to those skilled in the art, including ZF, MMSE, maximum likelihood, or MAP receivers. In one preferred embodiment, the receiver uses an MMSE filter to cancel ICI caused by IQ imbalance on the mirror tones. In another preferred embodiment, the receiver jointly detects symbols on the mirror tones using a nonlinear detector such as a maximum likelihood search. This method improves performance at the expense of increased complexity.

[0201] In one embodiment, the IQ-aware channel estimator 1204 is used to determine receiver coefficients that cancel ICI. Consequently, the inventors claim a DIDO-OFDM system that uses precoding based on channel state information, an IQ-aware DIDO receiver, and an IQ-aware channel estimator to cancel inter-carrier interference (ICI) due to mirror tones (due to I / Q mismatch). The channel estimator can use conventional training signals or specially constructed training signals sent with in-phase and quadrature signals. Any number of estimation algorithms can be implemented, including least squares, MMSE, or maximum likelihood. The IQ-aware channel estimator provides the input of the IQ-aware receiver.

[0202] Channel state information can be provided to the station through channel interaction or through a feedback channel. One embodiment of the present invention includes a DIDO-OFDM system with an I / Q-aware precoder and an I / Q-aware feedback channel that conveys channel state information from the user terminal to the station. The feedback channel can be a physical or logical control channel. It can be dedicated or shared, as in the case of a random access channel. The feedback information can be generated using a DIDO feedback generator at the user terminal, which is also claimed. The DIDO feedback generator takes the output of an I / Q-aware channel estimator as input. The channel coefficients can be quantized, or any of a limited number of feedback algorithms known in the art can be used.

[0203] The allocation of users, their modulation and coding rates, and their mapping to space-time-frequency code slots may change as a result of the DIDO feedback generator. Thus, one embodiment includes an IQ-aware DIDO composer that uses IQ-aware channel estimates from one or more users to form a DIDO IQ-aware precoder, i.e., selects the modulation rate, coding rate, subset of users that are allowed to transmit, and their mapping to space-time-frequency code slots.

[0204] To evaluate the performance of the proposed correction method, three DIDO2x2 systems are compared. 1. For I / Q mismatch, transmit on all tones (except DC and edge tones) without compensation for I / Q mismatch. 2. For I / Q compensation, transmit on all tones and compensate for I / Q mismatch by using "Method 1" above. 3. Ideal: Transmit only on odd tones to avoid inter-user and inter-carrier interference caused by I / Q mismatch (ie due to mirror tones).

[0205] Hereafter, we present results obtained from measurements on a DIDO-OFDM prototype in a real propagation scenario. Figure 14 shows the 64-QAM constellations obtained from the three systems mentioned above. These constellations are obtained at the same user location and with a constant average signal-to-noise ratio (-45 dB). The first constellation 1401 is very noisy due to interference from mirror tones caused by I / Q imbalance. The second constellation 1402 shows some improvement due to I / Q correction. The second constellation 1402 is not as clean as the ideal case shown as constellation 1403 due to the possibility of phase noise causing inter-carrier interference (ICI).

[0206] Figure 15 shows the average SER (Symbol Error Rate) 1501 and per-user goodput 1502 performance of a 2x2 DIDO system with 64-QAM and coding rate 3 / 4 with and without I / Q mismatch. The OFDM bandwidth is 250 kHz, the number of tones is 64, and the cyclic prefix length is L cp = 4. Since in the ideal case we transmit data only on a subset of the tones, the SER and goodput performance are evaluated as a function of the average per-tone transmit power (rather than the total transmit power) to ensure a fair comparison across different cases. Furthermore, in the following results, we use normalized values ​​of transmit power (expressed in decibels) because our goal is to compare the relative (rather than absolute) performance of different schemes. Figure 15 shows that in the presence of I / Q imbalance, the SER is significantly lower than the target SER (~10 -2 ) is shown to saturate before reaching 0. This saturation effect is due to the fact that the signal and interference (from mirror tones) power increases when the TX power increases. However, through the proposed I / Q compensation method, the interference can be canceled out to obtain better SER performance. Note that the slight increase in SER at high SNR is due to the amplitude saturation effect in the DAC due to the increased transmit power required for 64-QAM modulation.

[0207] Furthermore, observe that the SER performance with I / Q correction is very close to the ideal case. The 2 dB difference in TX power between these two cases is due to the possibility of phase noise causing additive interference between adjacent OFDM tones. Finally, the goodput curve 1502 shows that twice as much data can be transmitted when the I / Q method is applied compared to the ideal case because all data tones are used instead of the odd tones (as in the ideal case).

[0208] Figure 16 shows a graph of the SER performance of different QAM constellations with and without I / Q compensation. In this embodiment, the inventors find the proposed method particularly useful for 64-QAM constellations. For 4-QAM and 16-QAM, the present method of I / Q compensation results in worse performance than the case with I / Q mismatch, in some cases because the proposed method requires increased power to enable both data transmission and interference cancellation due to mirror tones. Furthermore, 4-QAM and 16-QAM are less affected by I / Q mismatch than 64-QAM due to the larger minimum distance between constellation points. See A. Tarighat, R. Bagheri, and A. H. Sayed, "Compensation Scheme and Performance Analysis of I / Q Imbalance in OFDM Receivers," IEEE Transactions on Signal Processing (see also IEEE Transactions on Acoustics, Speech, and Signal Processing), Vol. 53, pp. 3257-3268, August 2005. This can also be seen in Figure 16 by comparing the I / Q mismatch with the ideal case for 4-QAM and 16-QAM. Therefore, the additional power required by the DIDO precoder for interference cancellation (due to mirror tones) does not justify the small benefit of I / Q compensation for the 4-QAM and 16-QAM cases. Note that this problem can be solved by using methods 2 and 3 for I / Q compensation described above.

[0209] Finally, the relative SER performance of the above three methods is measured under different propagation conditions. For reference, the SER performance with I / Q mismatch is also illustrated. Figure 17 shows the measured SER for a DIDO 2x2 system with 64-QAM at two different user locations with a carrier frequency of 450.5 MHz and a bandwidth of 250 kHz. At location 1, the user is located in a different room, ~6λ from the BS, and is in NLOS (non-line-of-sight) conditions. At location 2, the user is located ~6λ from the BS and is in LOS (line-of-sight).

[0210] Figure 17 shows that not all three compensation methods necessarily perform better than the uncompensated case. Furthermore, Method 3 outperforms the other two compensation methods in any channel scenario. It should be noted that the relative results of Methods 1 and 2 depend on the propagation conditions. Practical measurement work shows that Method 1 usually performs better than Method 2 because it pre-cancels (at the transmitter) the inter-user interference caused by I / Q imbalance. When this inter-user interference is minimal, Method 2 can outperform Method 1, as shown in graph 1702 of Figure 17, because there is no power loss due to the I / Q compensation precoder.

[0211] Up until now, the comparison of different methods has been performed by considering only a limited set of propagation scenarios, as in the case of Figure 17. From now on, the relative results of these methods in an ideal iid (independent and identically distributed) channel will be measured. A DIDO-OFDM system provides simulations with I / Q phase and gain imbalance at the transmit and receive sides. Figure 18 shows the results of the proposed method for gain imbalance at the transmit side only (i.e., a gain of 0.8 on one rail and a gain of 1 on the other rail of the first transmit chain). It can be seen that Method 3 outperforms all other methods. Also, Method 1 performs better than Method 2 in an iid channel, in contrast to the results obtained at the two positions of graph 1702 in Figure 17.

[0212] That is, considering the three novel methods for correcting I / Q imbalance in a DIDO-OFDM system described above, Method 3 outperforms the other proposed correction methods. In systems with slow feedback channels, Method 2 can be used to reduce the amount of feedback required for the DIDO precoder at the expense of worse SER results.

[0213] II. Adaptive DIDO transmission method Another embodiment of a system and method for improving the performance of a distributed input distributed output (DIDO) system is described herein. The method dynamically allocates radio resources to different user devices by tracking changing channel conditions to meet a certain target error rate while increasing throughput. The user devices estimate and feed back their channel qualities to a base station. The base station processes the channel qualities obtained from the user devices and selects the best combination of user devices, DIDO scheme, modulation / coding scheme (MCS), and array configuration for the next transmission. The base station transmits parallel data to multiple user devices through precoding, and the signals are demodulated at the receiver.

[0214] A system for efficiently allocating resources to DIDO radio links is also described, including a DIDO base station having a DIDO configurer that processes feedback received from users to select the best set of users, DIDO schemes, modulation / coding schemes (MCSs), and array configurations for the next transmission, a receiver in the DIDO system that measures channel and other related parameters to generate a DIDO feedback signal, and a DIDO feedback control channel that conveys feedback information from users to the base station.

[0215] As will be described in detail below, some of the significant features of this embodiment of the present invention include, but are not limited to:

[0216] A technique for adaptively selecting a number of users, DIDO transmission scheme (i.e., antenna selection or multiplexing), modulation / coding scheme (MCS), and array configuration based on channel quality information in order to minimize SER or maximize per-user or downlink spectral efficiency.

[0217] A technique for defining a set of DIDO transmission modes as a combination of DIDO scheme and MCS.

[0218] A technique that assigns different DIDO modes to different time slots, OFDM tones, and DIDO substreams depending on the channel conditions.

[0219] A technique that dynamically allocates different DIDO modes to different users based on channel quality.

[0220] Criteria that enable adaptive DIDO switching based on link quality metrics calculated in the time, frequency, and spatial domains.

[0221] Criteria that enable adaptive DIDO switching based on a lookup table.

[0222] A DIDO system with a DIDO configurer at the base station, as in Figure 19, that adaptively selects a number of users, DIDO transmission scheme (i.e., antenna selection or multiplexing), modulation / coding scheme (MCS), and array configuration based on channel quality information to minimize SER or maximize per-user or downlink spectral efficiency.

[0223] A DIDO system having a DIDO configurer at the base station and a DIDO feedback generator at each user device, as in Figure 20, which uses estimated channel conditions and / or other estimated parameters such as SNR at the receiver to generate feedback messages that are input to the DIDO configurer.

[0224] A DIDO system having a DIDO configurator at a base station, a DIDO feedback generator, and a DIDO feedback control channel that conveys DIDO-specific configuration information from users to the base station.

[0225] a. background Within multiple-input multiple-output (MIMO) systems, diversity techniques such as orthogonal space-time block coding (OSTBC) (see V. Tarokh, H. Jafarkhani, and A. C. Calderbank, "Space-Time Block Codes with Orthogonal Constellations," IEEE Transactions on Information Theory, Vol. 45, pp. 1456-467, July 1999) or antenna selection (see R. W. Heath, Jr., S. Sandhu, and A. J. Paulraj, "Antenna Selection for Spatial Multiplexing Systems with Linear Receivers," IEEE Transactions on Communications, Vol. 5, pp. 142-144, April 2001) are expected to increase link robustness, combating channel fading and improving coverage. Meanwhile, spatial multiplexing (SM) allows the transmission of multiple parallel data streams as a means of improving system throughput. See GJ Foschini, GD Golden, RA Valenzuela, and PW Wolniansky, "Simplified Processing for High Spectral Efficiency Wireless Communications Using Multi-Element Arrays," Proceedings of the IEEE Society on Selected Areas of Communications, Vol. 17, No. 11, pp. 1841-1852, November 1999. These benefits can be simultaneously achieved in MIMO systems according to the theoretical diversity / multiplexing tradeoff derived in L. Zheng and D.N.C.Tse, "Diversity and Multiplexing: Fundamental Tradeoffs in Multi-Antenna Channels," IEEE Transactions on Information Theory, Vol. 49, No. 5, pp. 1073-1096, May 2003. One practical implementation is to adaptively switch between diversity and multiplexing transmission schemes by tracking changing channel conditions.

[0226] Several adaptive MIMO transmission techniques have been proposed. R.W. Heath, Jr., S. Sandhu, and A.J. Paulraj, "Switching Between Diversity and Multiplexing in MIMO Systems," IEEE Transactions on Communications, Vol. 53, No. 6, pp. 962-968, June 2005, proposes a diversity / multiplexing switching method that improves the bit error rate (BER) for transmission at a constant rate based on instantaneous channel quality information. Alternatively, statistical channel information can be used to enable adaptation, as in S. Catreux, V. Erceg, D. Gesbert, and R.W. Heath, Jr., "Adaptive Modulation and MIMO Coding for Broadband Wireless Data Communication Networks," IEEE Communications Journal, Vol. 2, pp. 108-115, June 2002 (Catreux), resulting in reduced feedback overhead and the number of control messages. Catreux's adaptive transmission algorithm was designed to improve spectral efficiency for a given target error rate in orthogonal frequency division multiplexing (OFDM) systems based on a channel time / frequency selectivity index. A similar low-feedback adaptive approach that exploits channel spatial selectivity by switching between diversity and spatial multiplexing has been proposed for narrowband systems.For example, "Adaptive MIMO Transmission Exploiting Spatially Correlated Channel Capabilities" by A. Forenza, M. R. McKay, A. Pandharipande, R. W. Heath Jr., and I. B. Collings, "Multiplexing / Beamforming Switching for Coded MIMO in Spatially Correlated Rayleigh Channels" by M. R. McKay, I. B. Collings, A. Forenza, and R. W. Heath Jr., "Switching Between OSTBC and Spatial Multiplexing with a Linear Receiver in Spatially Correlated MIMO Channels" by A. Forenza, M. R. McKay, R. W. Heath Jr., and I. B. Collings, "Vehicle Transmitters: Adaptive MIMO Transmission Exploiting Spatially Correlated Channel Capabilities ... Conf.,” IEEE Transactions on Cognitive and Signal Processing, Vol. 3, pp. 1387-1391, May 2006; see M.R. McKay, I.B. Collings, A. Forenza, and R.W. Heath, Jr., “Throughput-Based Adaptive MIMO-BICM Techniques for Spatially Correlated Channels,” to be published in IEEE-ICC Transactions on Cognitive and Signal Processing, June 2006.

[0227] Here, we extend the scope of work published in various previous papers to DIDO-OFDM systems, e.g., R.W. Heath, Jr., S. Sandhu, and A.J. Paulraj, "Switching Between Diversity and Multiplexing in MIMO Systems," IEEE Transactions on Communications, Vol. 53, No. 6, pp. 962-968, June 2005; S. Catreux, V. Erceg, D. Gesbert, and R.W. Heath, Jr., "Adaptive Modulation and MIMO Coding for Broadband Wireless Data Communication Networks," IEEE Communications Magazine, Vol. 2, pp. 108-115, June 2002 (Catreux); A. Forenza, M.R. McKay, A. Pandharipande, R.W. Heath, Jr., and I.B. Collings, "Spatial Phase Shift Keying (SPCE)," IEEE Transactions on Communications, Vol. 1, pp. 108-115, June 2002 (Catreux); "Adaptive MIMO Transmission Exploiting Related Channel Features," IEEE Transactions on Veh. Tech., Vol. 56, No. 2, pp. 619-630, March 2007. M.R. McKay, I.B. Collings, A. Forenza, and R.W. Heath, Jr., "Multiplexing / Beamforming Switching for Coded MIMO in Spatially Correlated Rayleigh Channels," A. Forenza, M.R. McKay, R.W. Heath, Jr., and I.B. Collings, "Switching Between OSTBC and Spatial Multiplexing with a Linear Receiver in Spatially Correlated MIMO Channels," Veh. Technol., Vol. 56, No. 2, pp. 619-630, March 2007, accepted for publication in IEEE Transactions on Veh. Tech., December 2007. Conf.,” IEEE Transactions on Cognitive and Signal Processing, Vol. 3, pp. 1387-1391, May 2006; see M.R. McKay, I.B. Collings, A. Forenza, and R.W. Heath, Jr., “Throughput-Based Adaptive MIMO-BICM Techniques for Spatially Correlated Channels,” to be published in IEEE-ICC Transactions on Cognitive and Signal Processing, June 2006.

[0228] This paper describes a novel adaptive DIDO transmission strategy that switches between different numbers of users, transmit antennas, and transmission schemes based on channel quality information as a means of improving system performance. Note that adaptive user selection within a multi-user MIMO system has been proposed in M. Sharif and B. Hassibi, "On the Performance of MIMO Broadcast Channels with Partial Side Information," IEEE Transactions on Information Theory, Vol. 51, pp. 506-522, February 2005, and W. Choi, A. Forenza, J.G. Andrews, and R.W. Heath, Jr., "Opportunistic Space Division Multiple Access Schemes with Beam Selection," to be published in IEEE Transactions on Communications. However, the opportunistic space division multiple access (OSDMA) schemes in these publications are designed to maximize sum function by exploiting multi-user diversity, and achieve only a fraction of the theoretical performance of dirty-paper codes because they do not fully pre-cancel interference at the transmitter. In the DIDO transmission algorithm described here, block diagonalization is used to pre-cancel inter-user interference, but the proposed adaptive transmission strategy can be applied to any DIDO system, independent of the type of precoding method.

[0229] This patent application describes extensions of the embodiments of the invention described above and in the prior application to include, but not limited to, the following additional features. 1. Prior Application of Channel Estimation Training symbols can be used by wireless client devices to estimate link quality metrics in an adaptive DIDO scheme. 2. The base station receives signal characterization data from the client device as described in the prior application. In this embodiment, the signal characterization data is defined as a link quality metric that is used to enable adaptation. 3. The prior application describes a mechanism for selecting transmit antennas and the number of users, as well as defining throughput allocation. Furthermore, different levels of throughput can be dynamically allocated to different clients, as in the prior application. In this embodiment of the present invention, we define new criteria related to this selection and throughput allocation.

[0230] b. Embodiments of the present invention The goal of the proposed adaptive DIDO method is to improve per-user or downlink spectral efficiency by dynamically allocating radio resources in time, frequency, and space to different users in the system. The general adaptation criterion is to increase throughput while meeting a target error rate. Based on propagation conditions, the adaptive algorithm can be used to improve the link quality of a user (or coverage) through diversity techniques. The flow chart shown in Figure 21 illustrates the steps of the adaptive DIDO method.

[0231] A base station (BS) collects channel state information (CSI) from all users in 2102. From the received CSI, the BS calculates link quality metrics in the time / frequency / spatial domain in 2104. These link quality metrics are used to select the users to be served in the next transmission as well as the transmission mode for each of the users in 2106. Note that the transmission modes consist of different combinations of modulation / coding and DIDO schemes. Finally, the BS transmits data to the users through DIDO precoding as in 2108.

[0232] In 2102, the base station collects channel state information (CSI) from all user devices. The CSI is used by the base station to determine instantaneous or statistical channel quality for all user devices in 2104. In a DIDO-OFDM system, channel quality (or link quality metrics) can be estimated in the time, frequency, and spatial domains. Then, in 2106, the base station uses the link quality metrics to determine the best subset of users and transmission mode for the current propagation conditions. A set of DIDO transmission modes is defined as a combination of DIDO schemes (i.e., antenna selection or multiplexing), modulation / coding schemes (MCS), and array configurations. In 2108, data is transmitted to the user devices using the selected number of users and transmission modes.

[0233] Mode selection is enabled by pre-calculated look-up tables (LUTs) based on the error rate performance of the DIDO system in different propagation environments. These LUTs map channel quality information to error rate performance. To construct the LUTs, the error rate performance of the DIDO system is evaluated in different propagation scenarios as a function of SNR. From the error rate curves, the minimum SNR required to achieve a certain predetermined target error rate can be calculated. This SNR requirement is defined as the SNR threshold. The SNR threshold is then evaluated in different propagation scenarios and for different DIDO transmission modes and stored in the LUTs. For example, the LUTs can be constructed using the SER results in Figures 24 and 26. From the LUTs, the base station then selects a transmission mode for the active user that meets the predetermined target error rate while simultaneously increasing throughput. Finally, the base station transmits data to the selected user through DIDO precoding. Note that different time slots, OFDM tones, and DIDO substreams can be assigned different DIDO modes so that adaptation can be performed in the time, frequency, and spatial domains.

[0234] One embodiment of a system using DIDO adaptation is shown in Figures 19-20. Several new functional units are introduced to enable implementation of the proposed DIDO adaptation algorithm. Specifically, in one embodiment, a DIDO configurator 1910 performs several functions, including selecting the number of users, the DIDO transmission scheme (i.e., antenna selection or multiplexing), the modulation / coding scheme (MCS), and the array configuration based on channel quality information 1912 provided by the user equipment.

[0235] The user selector unit 1902 selects the users U1-U2 based on the feedback information obtained by the DIDO configurer 1910. M The coding and modulation unit 1904 selects data associated with each user and provides this information to each of a plurality of coding and modulation units 1904. Each coding and modulation unit 1904 codes and modulates information bits for each user and sends them to a mapping unit 1906. The mapping unit 1906 maps the input bits to complex symbols and sends them to a precoding unit 1908. The coding and modulation units 1904 and mapping unit 1906 select the type of modulation / coding scheme to use for each user using information obtained from a DIDO configuration unit 1910. This information is calculated by the DIDO configuration unit 1910 by using each user's channel quality information provided by a feedback unit 1912. The DIDO precoding unit 1908 uses the information obtained by the DIDO configuration unit 1910 to calculate DIDO precoding weights and precode the input symbols obtained from the mapping unit 1906. Each of the precoded data streams is sent by the DIDO precoding unit 1908 to an OFDM unit 1915, which calculates an IFFT and adds a cyclic prefix. This information is sent to a D / A unit 1916 which performs a digital-to-analog conversion and sends the resulting analog signal to an RF unit 1914. The RF unit 1914 upconverts the baseband signal to a mid / high frequency and sends it to a transmit antenna.

[0236] The RF unit 2008 of each client device receives the signal transmitted from the DIDO transmitter unit 1914, downconverts the signal to baseband, and provides the downconverted signal to the A / D unit 2010. The A / D unit 2010 then converts the signal from analog to digital and sends it to the OFDM unit 2013. The OFDM unit 2013 removes the cyclic prefix and performs an FFT to translate the signal into the frequency domain. During the training period, the OFDM unit 2013 sends its output to the channel estimator 2004, which calculates a channel estimate in the frequency domain. Alternatively, the channel estimate can be calculated in the time domain. During the data period, the OFDM unit 2013 sends its output to the receiver unit 2002, which demodulates / decodes the signal to obtain the data 2014. The channel estimator 2004 sends the channel estimate to the DIDO feedback generator 2006, which can quantize the channel estimate and send it to the transmitter over the feedback control channel 1912.

[0237] The DIDO configurer 1910 may use information derived at the base station, or in a preferred embodiment, further uses the output of a DIDO feedback generator 2006 (see FIG. 20) operating at each user device. The DIDO feedback generator 2006 uses estimated channel conditions 2004 and / or other parameters, such as SNR estimated at the receiver, to generate feedback messages to be input to the DIDO configurer 1910. The DIDO feedback generator 2006 may compress the information at the receiver, quantize the information, and / or use any limited feedback strategy known in the art.

[0238] The DIDO configurator 1910 can use information recovered from the DIDO feedback control channel 1912, which is a logical or physical control channel used to send the output of the DIDO feedback generator 2006 from the user to the base station. The control channel 1912 can be a logical or physical control channel that can be implemented in any number of ways known in the art. As a physical channel, it can include dedicated time / frequency slots assigned to users. It can be a random access channel shared by all users. The control channel can be pre-assigned or can be created by stealing bits in a predetermined manner from an existing control channel.

[0239] In the following discussion, we describe results obtained through measurements with a DIDO-OFDM prototype in a real-world propagation environment. These results highlight the potential gains achievable in an adaptive DIDO system. We first demonstrate the performance of a DIDO system with different orders, demonstrating the ability to increase the number of antennas / users to provide increased downlink throughput. We then discuss DIDO performance as a function of user device location, highlighting the need to track changing channel conditions. Finally, we discuss the performance of a DIDO system using diversity techniques.

[0240] i. Performance of the different-order DIDO system The performance of different DIDO systems is evaluated with increasing number of transmit antennas N=M, where M is the number of users. We compare the performance of the following systems: SISO, DIDO2x2, DIDO4x4, DIDO6x6, and DIDO8x8. DIDONxM refers to DIDO with N transmit antennas at the BS and M users.

[0241] Figure 22 shows the transmit / receive antenna arrangement. The transmit antennas 2201 are installed in a rectangular array configuration, and users are located around the transmit array. In Figure 22, T denotes a "transmit" antenna, and U denotes a "user device" 2202.

[0242] Different antenna subsets are active in the 8-element transmit array depending on the value of N selected for different measurements. For each DIDO order (N), we selected the subset of antennas that covers the maximum real estate within a given size constraint of the 8-element array. This criterion is expected to improve spatial diversity for any given value of N.

[0243] Figure 23 shows array configurations for different DIDO orders that fit into the available real estate (i.e., dashed lines). The rectangular dashed box has dimensions of 24"x24", corresponding to ~λxλ at a carrier frequency of 450 MHz.

[0244] Based on the notes associated with FIG. 23, and with reference to FIG. 22, the performance of each of the following systems will now be defined and compared. SISO(2301) with T1 and U1 DIDO2x2 (2302) with T1, 2 and U1, 2 DIDO 4x4 (2303) by T1, 2, 3, 4 and U1, 2, 3, 4 DIDO 6x6 (2304) by T1, 2, 3, 4, 5, 6 and U1, 2, 3, 4, 5, 6 DIDO 8x8 (2305) with T1, 2, 3, 4, 5, 6, 7, 8 and U1, 2, 3, 4, 5, 6, 7, 8

[0245] Figure 24 shows the SER, BER, SE (spectral efficiency), and goodput performance as a function of transmit (TX) power for the above-described DIDO system with 4-QAM and a FEC (forward error correction) rate of 1 / 2. Observe that the SER and BER performance deteriorate as the value of N increases. This effect is due to two phenomena: for a constant TX power, the input power to the DIDO array is divided among an increasing number of users (or data streams), and spatial diversity decreases with an increasing number of users in a real (spatially correlated) DIDO channel.

[0246] To compare the relative performance of different-order DIDO systems, the target BER is set to SER=10 as shown in Figure 24. -2 10, which roughly corresponds to -4 (This value may vary based on the system). The TX power value corresponding to this target is called the TX power threshold (TPT). For any N, if the TX power is less than TPT, we assume that it is not possible to transmit with DIDO order N, and we need to switch to a lower-order DIDO. Also, observe in Figure 24 that the SE and goodput performance saturates when the TX power exceeds TPT for any value of N. From these results, we can devise an adaptive transmission strategy that switches between different-order DIDO to improve SE or goodput so that a certain predetermined target error rate is obtained.

[0247] ii. Performance with various user locations The goal of this experiment is to evaluate DIDO performance for different user positions through simulations in a spatially correlated channel. A DIDO 2x2 system is considered with 4QAM and an FEC rate of 1 / 2. User 1 is located broadside from the transmit array, while user 2 is moved from broadside to endfire, as shown in Figure 25. The transmit antennas are spaced ~λ / 2 apart and separated from the users by ~2.5λ.

[0248] Figure 26 shows the SER and per-user SE results for different positions of two user devices. The angle of arrival (AOA) of the user devices ranges between 0° and 90° measured from the broadside direction of the transmit array. Observe that as the separation angle of the user devices increases, the DIDO performance improves due to the increased diversity available in the DIDO channel. Also, for a target SER = 10 -2 , there is a 10 dB gap between the cases AOA2=0° and AOA2=90°. This result is in agreement with the simulation results obtained in Figure 35 for an angular spread of 10°. Also, note that for the case AOA1=AOA2=0° there may be coupling effects between the two users (due to the proximity of the antennas) which may cause the performance to vary from the simulated results in Figure 35.

[0249] iii. Favorable scenario for DIDO8x8 Figure 24 shows that 8x8 DIDO provides the SE increase due to the lower order of DIDO at the cost of higher TX power requirements. The goal of this experiment is to show that there are cases where DIDO8x8 outperforms DIDO2x2 not only in terms of peak spectral efficiency (SE), but also in terms of the TX power requirement (or TPT) that results in that peak SE.

[0250] Note that there is a ~6 dB gap in TX power between the SEs of DIDO2x2 and DIDO8x8 in an iid (ideal) channel. This gap is due to the fact that DIDO8x8 divides the TX power across eight data streams, while DIDO2x2 divides the TX power between only two streams. This result is shown through simulation in Figure 32.

[0251] However, in spatially correlated channels, TPT is a function of the properties of the propagation environment (e.g., array orientation, user position, angular spread). For example, Figure 35 shows a ~15 dB difference for low angular spread for two different user device locations. Similar results are shown in Figure 26 of this application.

[0252] Similar to MIMO systems, the performance of DIDO systems deteriorates when users are located in the endfire direction from the TX array (due to the lack of diversity). This effect has been observed through measurements on current DIDO prototypes. Therefore, one way to demonstrate the case where DIDO8x8 outperforms DIDO2x2 is to place a user in the endfire direction relative to the DIDO2x2 array. In this scenario, DIDO8x8 outperforms DIDO2x2 due to the higher diversity achieved by the eight-antenna array.

[0253] In this analysis, the following system is considered:

[0254] System 1: DIDO8x8 with 4-QAM (transmitting eight parallel data streams per time slot).

[0255] System 2: DIDO2x2 with 64-QAM (transmitting to users X and Y every 4 time slots). For this system, four combinations of TX and RX antenna positions are considered: a) T1, T2, U1,2 (endfire direction), b) T3, T4, U3,4 (endfire direction), c) T5, T6, U5,6 (~30° from endfire direction), d) T7, T8, U7,8 (NLO (non-line-of-sight)).

[0256] System 3: DIDO 8x8 with 64-QAM.

[0257] System 4: MISO8x1 with 64-QAM (transmit to user X every 8 time slots).

[0258] For all these cases, an FEC rate of 3 / 4 was used.

[0259] The user's position is shown in FIG.

[0260] In Figure 28, the SER results show a ~15 dB gap between Systems 2a and 2c with different array orientations and user positions (similar to the simulation results in Figure 35). The first subplot in the second row shows the value of TX power at which the SE curve saturates (i.e., corresponding to a BER of 1e-4). It is observed that System 1 achieves an increase in per-user SE, resulting in a lower TX power requirement (~5 dB) than System 2. Also, the benefit of DIDO8x8 over DIDO2x2 is more evident in terms of D1 (downlink) SE and D1 goodput due to the multiplexing gain of DIDO8x8 over DIDO2x2. System 4 achieves a lower TX power requirement (less than 8 dB) than System 1 due to the array gain of beamforming (i.e., MRC with MISO8x1). However, System 4 achieves only one-third of the per-user SE compared to System 1. In System 2, performance is worse than in System 1 (i.e., the larger the TX power requirement, the smaller the SE). Finally, in System 3, the larger the TX power requirement (15 dB), the larger the SE is than in System 1.

[0261] From these results, the following conclusions can be drawn:

[0262] One channel scenarios were identified where DIDO8x8 outperformed DIDO2x2 (i.e., the higher the TX power requirement, the greater the SE).

[0263] In this channel scenario, the SE per user with DIDO8x8 and the SE of D1 are larger than those with DIDO2x2 and MISO8x1.

[0264] The performance of DIDO8x8 can be further increased by using a higher order of modulation (ie, 64-QAM instead of 4-QAM) at the cost of increased TX power requirements (>~15dB).

[0265] iv. DIDO with antenna selection Hereafter, we evaluate the benefits of the antenna selection algorithm described in "Transmit Selection Diversity for Single-Precoded Multiuser Spatial Multiplexing Systems with Linear Receivers" by R. Chen, R.W. Heath, and J.G. Andrews, accepted for publication in IEEE Transactions on Signal Processing, 2005. Results are shown for one particular DIDO system with two users, 4-QAM, and an FEC rate of 1 / 2. The following systems are compared in Figure 27: DIDO2x2 with T1,2 and U1,2, and DIDO3x2 with antenna selection by T1,2,3 and U1,2.

[0266] The positions of the transmitting antennas and user equipment are the same as in FIG.

[0267] Figure 29 shows that DIDO3x2 with antenna selection can achieve a gain of ~5 dB compared to a DIDO2x2 system (without selection). Note that the channel is nearly static (i.e., no Doppler), and therefore the selection algorithm adapts to path loss and channel spatial correlation rather than fast fading. Scenarios with high Doppler should see different gains. Also, in this particular experiment, the antenna selection algorithm was observed to select antennas 2 and 3 for transmission.

[0268] iv. SNR threshold for LUT In the "Method 2 - RX Correction" section, mode selection was described as being enabled by a LUT. The LUT can be pre-calculated by evaluating the SNR threshold to yield a certain predetermined target error rate performance for the DIDO transmission mode in different propagation environments. Hereafter, we present the performance of a DIDO system with and without antenna selection and with a variable number of users, which can be used as a guide to construct the LUT. Figures 24, 26, 28, and 29 are derived from practical measurements with a DIDO prototype, while the following figures are obtained through simulation. The following BER results assume no FEC.

[0269] Figure 30 shows the average BER performance of different DIDO precoding schemes in an iid channel. The curve labeled "No Selection" refers to the case when BD is used. In the figure, the performance of antenna selection (ASel) is shown for different numbers of extra antennas (relative to the number of users). It can be seen that as the number of extra antennas increases, ASel improves the diversity gain (characterized by the slope of the BER curve in the high SNR region) and therefore the reception range. For example, for 10 -2 For a fixed target BER (which is the actual value for an uncoded system), the SNR gain achieved by ASel increases with the number of antennas.

[0270] Figure 31 shows the SNR gain of ASel as a function of the number of extra transmit antennas in an iid channel for different target BERs. It can be seen that with just one or two additional antennas, ASel can achieve significant SNR gains compared to BD. In the following sections, we only consider the cases of one or two extra antennas and 10 -2 We evaluate the performance of ASel by fixing the target BER at (for uncoded systems).

[0271] Figure 32 shows the SNR threshold as a function of the number of users (M) for BD and ASel for one and two extra antennas in an iid channel. We observed that the SNR threshold increases with M because the required received SNR increases with a larger number of users. Note that we assume a constant total transmit power for all numbers of users (the number of transmit antennas is variable). Furthermore, Figure 32 shows that the gain from antenna selection is constant regardless of the number of users in an iid channel. Hereafter, we demonstrate the performance of DIDO systems in spatially correlated channels. We provide channel simulations for each user through the COST-259 spatial channel model described in X. Zhuang, FW Vook, KL Baum, TA Thomas, and M. Cudak, "Channel Models for Link and System-Level Simulations," IEEE 802.16 Broadband Wireless Access Working Group, September 2004. We generate a single cluster for each user. As a case study, we assume an NLO channel with an element spacing of 0.5 lambda (for a uniform linear array (ULA) transmitter). For the two-user system case, we simulated clusters with mean angles of arrival AOA1 and AOA2 for the first and second users, respectively. AOA is measured relative to the broadside direction of the ULA. When two or more users are in the system, the range [-φ m , φ m ], where we define:

[0272]

number

[0273] where k is the number of users and Δφ is the separation angle between the average AOAs of the users. m , φ mNote that the center of ] is at an angle of 0°, corresponding to the broadside direction of the ULA. Hereafter, we consider the BER performance of the DIDO system as a function of the channel angular spread (AS) and separation angle between users, using BD and ASel transmission schemes and different numbers of users.

[0274] Figure 33 shows the BER versus average SNR per user for two users located at the same angle with different values ​​of AS (i.e., AOA1 = AOA2 = 0° relative to the broadside direction of the ULA). It can be seen that the BER performance improves as AS increases, approaching the iid case. In fact, the higher the AS, the smaller the statistical overlap between the eigenmodes of the two users and the better the performance of the BD precoder.

[0275] Figure 34 shows similar results to Figure 33, but with a higher angular separation between users. Consider AOA1 = 0° and AOA2 = 90° (i.e., a 90° angular separation). The best performance is now achieved for the low AS case. In fact, for the large angular separation case, there is less overlap between user eigenmodes when the angular spread is low. Interestingly, we observe that the BER performance at low AS is better than for iid channels for the same reasons mentioned above.

[0276] Next, we compare the correlations in different scenarios. -2 We calculate the SNR threshold for a target BER of 1. Figure 35 plots the SNR threshold as a function of AS for different values ​​of the users' average AOA. When the users' angular separation is low, reliable transmission with adequate SNR requirements (i.e., 18 dB) is only possible for channels characterized by high AS. On the other hand, when users are spatially separated, a smaller SNR is required to meet the same target BER.

[0277] Figure 36 shows the SNR threshold for the five-user case. The average AOA of users is generated according to the definition in (13) for different values ​​of the separation angle Δφ. We observe that for Δφ = 0° and AS < 15°, BD performs poorly due to the small angular separation between users, and the target BER is not met. With increasing AS, the SNR requirement to meet a certain target BER decreases. On the other hand, for Δφ = 30°, the minimum SNR requirement is obtained at a low AS, consistent with the results in Figure 35. As AS increases, the SNR threshold saturates for one of the iid channels. Note that Δφ = 30° with five users corresponds to an AOA range of [-60°, 60°], which is typical for base stations in cellular systems with 120° sectorial cells.

[0278] Next, we consider the performance of the ASel transmission scheme in spatially correlated channels. Figure 37 compares the SNR thresholds of BD and ASel with one and two extra antennas for the two-user case, for example. Consider two different cases of inter-user separation angles: {AOA1=0°, AOA2=0°} and {AOA1=0°, AOA2=90°}. The curves for the BD scheme (i.e., without antenna selection) are the same as those in Figure 35. We observe that ASel can achieve SNR gains of 8 dB and 10 dB with one and two extra antennas, respectively, for high AS. As AS decreases, the gain from ASel in BD becomes smaller due to the reduced number of degrees of freedom in the MIMO broadcast channel. Interestingly, for AS=0° (i.e., near the LOS channel) and the case {AOA1=0°, AOA2=90°}, ASel does not provide any gain due to the lack of diversity in the spatial domain. Figure 38 shows similar results to Figure 37, but for the five user case.

[0279] For both BD and ASel transmission schemes, we calculate the SNR threshold (10 as a function of the number of users (M) in the system). -2(assuming a typical target BER of . The SNR threshold corresponds to the average SNR so that the total transmit power is constant for all M. The azimuth angle range [-φ m , φΔ m Assume a maximum separation between the average AOA of each user's cluster within [M, M] = [-60°, 60°]. Then, the angular separation between users is Δφ = 120° / (M-1).

[0280] Figure 39 shows the SNR threshold for the BD scheme with different values ​​of AS. We observe that the lowest SNR requirement is obtained for AS = 0.1° (i.e., low angular spread) with a relatively small number of users (i.e., K < 20) due to the large angular separation between users. However, for M > 50, the SNR requirement is much larger than 40 dB because Δφ is very small and BD is impractical. Furthermore, for AS > 10°, the SNR threshold remains nearly constant for all M, and the DIDO system in a spatially correlated channel approaches the performance of an iid channel.

[0281] To reduce the SNR threshold value and improve the performance of the DIDO system, we apply the ASel transmission scheme. Figure 40 shows the SNR threshold in spatially correlated channels at AS=0.1° for BD and ASel with one and two extra antennas. For reference, we also report the curves for the iid case shown in Figure 32. It can be observed that for low user numbers (i.e., M<10), antenna selection does not help reduce the SNR requirement due to the lack of diversity in the DIDO broadcast channel. As the number of users increases, ASel benefits from multi-user diversity and achieves an SNR gain (i.e., 4 dB for M=20). Furthermore, for M≦20, the performance of ASel with one or two antennas in a channel with high spatial correlation is the same.

[0282] Next, we calculate the SNR threshold for two more channel scenarios: AS=5° in Figure 41 and AS=10° in Figure 42. Figure 41 shows that ASel provides SNR gains even for a relatively small number of users (i.e., M≦10) due to the increased angular spread, in contrast to Figure 40. For AS=10°, the SNR threshold is further reduced as reported in Figure 42, and the gain with ASel is higher.

[0283] Finally, we summarize the results presented so far for correlated channels. Figures 43 and 44 show the SNR threshold as a function of the number of users (M) and angular spread (AS) for BD and ASel schemes with one and two extra antennas. The case of AS=30° actually corresponds to an iid channel, and this value of AS was used in the plots only for graphical representation. Although BD suffers from channel spatial correlation, we observe that ASel achieves almost the same performance for any AS. Furthermore, for AS=0.1°, ASel performs similarly to BD for low M, while for large M (i.e., M≧20), it outperforms BD due to multi-user diversity.

[0284] Figure 49 compares the performance of different DIDO schemes with respect to the SNR threshold. The DIDO schemes considered are BD, ASeI, and BD with eigenmode selection (BD-ESeI) and maximum ratio combining (MRC). Note that while MRC does not pre-cancel interference at the transmitter (unlike other methods), it does provide gain enhancement when users are spatially separated. Figure 49 plots a constant SNR threshold against a target BER of 10-2 for a DIDO Nx2 system when two users are located at -30° and 30° from the broadside direction of the transmit array, respectively. We observe that for low AS, the MRC scheme achieves a 3 dB gain compared to the other schemes because the users' spatial channels are well separated and the impact of inter-user interference is low. Note that the gain of MRC over DIDO Nx2 is due to the array gain. For ASs larger than 20°, the QR-ASel scheme outperforms the others, providing a 10 dB gain compared to BD2x2 without selection. QR-ASel and BD-ESeI provide the same performance for all values ​​of AS.

[0285] A new adaptive transmission technique for DIDO systems is described above. The method dynamically switches between DIDO transmission modes for different users to improve throughput for a constant target error rate. The performance of the different-order DIDO system is measured under different propagation conditions, and it is found that significant gains in throughput can be achieved by dynamically selecting the DIDO mode and number of users as a function of propagation conditions.

[0286] III. Frequency and Phase Offset Pre-Compensation a. background As described above, wireless communication systems communicate information using carrier waves. These carrier waves are typically sinusoids that are amplitude- and / or phase-modulated in response to the information to be transmitted. The nominal frequency of the sinusoids is known as the carrier frequency. To create this waveform, the transmitter provides upconversion to combine one or more sinusoids to create a modulated signal riding on the sinusoid at a predetermined carrier frequency. This can be done through direct conversion, where a signal is modulated directly onto a carrier or through multiple upconversion stages. To process this waveform, the receiver must demodulate the received RF signal to essentially cancel the modulated carrier. This requires the receiver to combine one or more sinusoidal signals to reverse the modulation scheme used by the transmitter, known as downconversion. Unfortunately, the sinusoidal signals generated by the transmitter and receiver are derived from different reference oscillators. No reference oscillator produces a perfect frequency reference; in practice, there will always be some deviation from the true frequency.

[0287] In wireless communication systems, differences in the power of the reference oscillators at the transmitter and receiver cause a phenomenon known as carrier frequency offset, or simply frequency offset, at the receiver. Essentially, some residual modulation (corresponding to differences in the transmitted and received carriers) is present in the received signal after downconversion. This causes distortion of the received signal, and therefore a high bit error rate and low throughput.

[0288] There are different ways to deal with carrier frequency offset: most approaches estimate the carrier frequency offset at the receiver and then apply a carrier frequency offset correction algorithm. Carrier frequency offset estimation algorithms have been developed based on the following characteristics: offset QAM (T. Fusco and M. Tanda, "Insensitive Frequency Offset Estimation for OFDM / OQAM Systems," IEEE Transactions on Signal Processing [see also IEEE Transactions on Acoustics, Speech, and Signal Processing], Vol. 55, pp. 1828-1838, 2007), periodic nature (E. Serpedin, A. Chevreuil, G.B. Giannakis, and P. Loubaton, "Insensitive Channel and Carrier Frequency Offset Estimation Using Periodic Modulation Precoders," IEEE Transactions on Signal Processing [see also IEEE Transactions on Acoustics, Speech, and Signal Processing], Vol. 48, No. 8, pp. 2389-2405, August 2000), or cyclic prefix in orthogonal frequency division multiplexing (OFDM) structural methods (J.J. van de Beek, M. Sandell, and P.O. Borjesson, "ML Estimation of Time and Frequency Offsets in OFDM Systems," IEEE Transactions on Signal Processing [see also IEEE Transactions on Acoustics, Speech, and Signal Processing], Vol. 45, No. 7, pp. 1800-1805, July 1997; U. Tureli, H. Liu, and M.D. Zoltowski, "OFDM-Blind Carrier Offset Estimation: ESPRIT," IEEE Transactions on Communications, Vol. 48, No. 9, pp. 1459-1461, September 2000; M. Luise, M. Marselli, and R. Reggiannini, "Low-Complexity Blind Carrier Frequency Recovery for OFDM Signals over Frequency-Selective Wireless Channels," IEEE Transactions on Communications, Vol. 50, No. 7, pp. 1182-1188, July 2002).

[0289] Alternatively, a special training signal containing repeated data symbols (P.H. Moose, "Orthogonal Frequency Division Multiplexing Frequency Offset Correction Method," IEEE Transactions on Communications, Vol. 42, No. 10, pp. 2908-2914, October 1994), two distinct symbols (T.M. Schmidl and D.C. Cox, "Robust Frequency and Timing Synchronization for OFDM," IEEE Transactions on Communications, Vol. 45, No. 12, pp. 1613-1621, December 1997), or a periodically inserted known symbol sequence (M. Luise, M. Marselli, and R. Reggiannini, "Carrier Frequency Acquisition and Tracking for OFDM Systems," IEEE Transactions on Communications, Vol. 44, No. 11, pp. 1590-1598, November 1996) can be used. Correction can be performed analog or digital. The receiver can precorrect the transmitted signal using the carrier frequency offset estimate to eliminate the offset. Carrier frequency offset correction has been widely considered for multicarrier and OFDM systems due to their sensitivity to frequency offsets (J.J. van de Beek, M. Sandell, and P.O. Borjesson, "ML Estimation of Time and Frequency Offsets in OFDM Systems," IEEE Transactions on Signal Processing [see also IEEE Transactions on Acoustics, Speech, and Signal Processing], Vol. 45, No. 7, pp. 1800-1805, July 1997; U. Tureli, H. Liu, and M.D. Zoltowski, "OFDM Insensitive Carrier Offset Estimation: ESPRIT," IEEE Transactions on Communications, Vol. 48, No. 9, pp. 1459-1469, July 1997). 1461-1461, September 2000; T.M. Schmidl and D.C. Cox, "Robust Frequency and Timing Synchronization for OFDM," IEEE Transactions on Communications, Vol. 45, No. 12, pp. 1613-1621, December 1997; M. Luise, M. Marselli, and R. Reggiannini, "Low-Complexity Insensitive Carrier Frequency Recovery for OFDM Signals over Frequency-Selective Wireless Channels," IEEE Transactions on Communications, Vol. 50, No. 7, pp. 1182-1188, July 2002.

[0290] Frequency offset estimation and correction is an important issue for multiple antenna communication systems, or more generally for MIMO (multiple input multiple output) systems. In a MIMO system, where the transmit antennas are locked to one frequency reference and the receiver is locked to another frequency reference, there is a single offset between the transmitter and the receiver. Several algorithms have been proposed to address this problem using training signals (K. Lee and J. Chun, "Frequency Offset Estimation for MIMO and OFDM Systems Using Orthogonal Training Sequences," IEEE Transactions on Veh. Technol., Vol. 56, No. 1, pp. 146-156, January 2007; M. Ghogho and A. Swami, "Training Design for Multipath Channel and Frequency Offset Estimation in MIMO Systems," IEEE Transactions on Signal Processing [see also IEEE Transactions on Acoustics, Speech, and Signal Processing], Vol. 54, No. 10, pp. 3957-3965, October 2005), and adaptive tracking in communications (C. Oberli and B. Daneshrad, "Maximum Likelihood Tracking Algorithm for MIMO and OFDM," 2004 IEEE International Conference on, Vol. 4, June 4-24, 2004, pp. 2468-2472). A more severe problem arises in MIMO systems where the transmit antennas are not locked to the same frequency reference, but the receive antennas are locked together. This effectively occurs in the uplink of a spatial division multiple access (SDMA) system, which can be viewed as a MIMO system where different users correspond to different transmit antennas. In this case, correcting for frequency offsets becomes much more complicated. In particular, frequency offsets cause interference between the different transmitted MIMO streams.This interference requires complex joint estimation and equalization algorithms (A. Kannan, T. P. Krauss, and M. D. Zoltowski, "Separation of Co-Channel Signals Under Imperfect Timing and Carrier Synchronization," IEEE Transactions on Veh. Technol., Vol. 50, No. 1, pp. 79-96, January 2001), and post-equalization frequency offset estimation (T. Tang and R. W. Heath, "Frequency Offset Estimation for MIMO-OFDM Systems [Mobile Radio]"). "Joint Frequency Offset Estimation and Interference Cancellation for OFDM / SDMA Systems," VTC Fall 2004, 2004 IEEE 60th Vehicle Technology Conference, Vol. 3, pp. 1553-1557, September 26-29, 2004; X. Dai, "Carrier Frequency Offset Estimation for OFDM / SDMA Systems Using Continuous Pilots," IEEE Transactions on Communications, Vol. 152, pp. 624-632, October 7, 2005. One study addressed the issue of residual phase offset and tracking error, where the residual phase offset is estimated and corrected after frequency offset estimation, but this study only considered the uplink of SDMA-OFDMA systems (L. Haring, S. Bieder, and A. Czylwik, "Residual Carrier and Sampling Frequency Synchronization in Multiuser OFDM Systems," 2006 VTC, Spring 2006, IEEE 63rd Vehicle Technology Conference, Vol. 4, pp. 1937-1941, 2006). The most severe case in MIMO systems occurs when all transmit and receive antennas have different frequency references. The only available work on this subject deals only with the asymptotic analysis of estimation errors in flat-fading channels (O. Besson and P. Stoica, "On Parameter Estimation for MIMO Flat-Fading Channels with Frequency Offsets," IEEE Transactions on Signal Processing [see also IEEE Transactions on Acoustics, Speech, and Signal Processing], Vol. 51, No. 3, pp. 602-613, March 2003).

[0291] A notably underexplored case occurs when the frequency references of different transmit antennas in a MIMO system are not the same and the receive antennas process the signals independently. This occurs in what are known as distributed input / distributed output (DIDO) communication systems (also referred to in the literature as MIMO broadcast channels). DIDO systems utilize the same radio resources (i.e., the same slot duration and frequency band) as traditional SISO systems, but consist of a single access point with distributed antennas transmitting parallel data streams (through precoding) to multiple users to improve downlink throughput. A detailed description of a DIDO system is provided in U.S. Patent Application Publication No. 20060023803, entitled "Distributed Input / Distributed Output Wireless Communication System and Method," by S.G. Perlman and T. Cotter, published in July 2004. There are many ways to implement a DIDO precoder.One solution is described, for example, in QH Spencer, ALS Windlehurst, and M. Haardt, "Zero-Forcing Methods for Downlink Spatial Multiplexing in Multiuser MIMO Channels," IEEE Transactions on Signal Processing, Vol. 52, pp. 461-471, February 2004; K.K. Wong, R.D. Murch, and K.B. Letaief, "Joint Channel Diagonalization for Multiuser MIMO Antenna Systems," IEEE Transactions on Wireless Communications, Vol. 2, pp. 773-786, July 2003; L.U. Choi and R.D. Murch, "Transmit Preprocessing Methods for Multiuser MIMO Systems Using Decomposition Techniques," IEEE Transactions on Wireless Communications, Vol. 3, pp. 20-24, January 2004; Z. Shen, J.G. Andrews, and R.W.H. Heath, and B. L. vans, "A Low-Complexity User Selection Algorithm for Multiuser MIMO Systems with Block Diagonalization," IEEE Transactions on Signal Processing, September 2005, accepted for publication; Z. Shen, R. Chen, J. G. Andrews, R. W. Heath, and B. L. vans, "Sum Function of Multiuser MIMO Broadcast Channels with Block Diagonalization," submitted to IEEE Transactions on Wireless Communications, October 2005; and R. Chen, R. W. Heath, and J. G. Andrews, "Transmit Selection Diversity for Single-Precoded Multiuser Spatial Multiplexing Systems with Linear Receivers," accepted for publication to IEEE Transactions on Signal Processing, 2005.

[0292] In DIDO systems, transmit precoding is used to separate data streams intended for different users. Carrier frequency offsets cause several system performance problems when the transmit antenna radio frequency chains do not share the same frequency reference. When this occurs, each antenna effectively transmits at a slightly different carrier frequency, destroying the integrity of the DIDO precoder and thus causing each user to experience special interference. Several solutions to this problem are proposed below. In one solution embodiment, DIDO transmit antennas share a frequency reference over a wired, optical, or wireless network. In another solution embodiment, one or more users estimate the frequency offset difference (the relative difference in offset between antenna pairs) and send this information to the transmitter. The transmitter then pre-compensates for the frequency offset and proceeds with the training and precoder estimation phases for DIDO. This embodiment presents a problem when there is a delay in the feedback channel. The reason is that in subsequent channel estimation, there may be residual phase errors caused by undescribed compensation methods. To solve this problem, one additional embodiment uses a new frequency offset and phase estimator that can correct this problem by estimating the delay. Results are presented based on simulations and practical measurements performed on a DIDO-OFDM prototype.

[0293] The frequency and phase offset correction methods proposed herein may be susceptible to estimation errors due to noise at the receiver. Therefore, in one additional embodiment, we propose a time and frequency offset estimation method that is robust even under low SNR conditions.

[0294] Different approaches exist for time and frequency offset estimation, and many of these approaches have been proposed specifically for OFDM waveforms due to their susceptibility to synchronization errors.

[0295] The algorithms typically do not exploit the structure of OFDM waveforms and are therefore general enough for both single-carrier and multi-carrier waveforms. The algorithm described below is one of a class of techniques that use known reference symbols, e.g., training data, to aid synchronization. Most of these methods are extensions of Moose's frequency offset estimator (see P.H. Moose, "Orthogonal Frequency Division Multiplexing Frequency Offset Correction Method," IEEE Transactions on Communications, Vol. 42, No. 10, pp. 2908-2914, October 1994). Moose proposed using two repeated training signals and derived the frequency offset using the phase difference between both received signals. Moose's method can only correct fractional frequency offset. An extension of the Moose method has been proposed by Schmidl and Cox (T.M. Schmidl and D.C. Cox, "Robust Frequency and Timing Synchronization for OFDM," IEEE Transactions on Communications, Vol. 45, No. 12, pp. 1613-1621, December 1997). Their key innovation was the use of one periodic OFDM symbol with an additional differentially coded training symbol. Differential coding within the second symbol allows for integer offset correction. Coulson considered similar configurations as described in T.M. Schmidl and D.C. Cox, "Robust Frequency and Timing Synchronization for OFDM," IEEE Transactions on Communications, Vol. 45, No. 12, pp. 1613-1621, December 1997, and detailed the algorithm and analysis as described in A.J. Coulson, "Maximum Likelihood Synchronization for OFDM Using Pilot Symbols: Analysis," IEEE Journal on Selected Areas of Communications, Vol. 19, No. 12, pp. 2495-2503, December 2001, and A.J. Coulson, "Maximum Likelihood Synchronization for OFDM Using Pilot Symbols: Algorithms," IEEE Journal on Selected Areas of Communications, Vol. 19, No. 12, pp. 2486-2494, December 2001.One major difference is that Coulson used repeated maximum-length sequences to obtain good correlation performance and suggested the use of chirp signals due to their constant envelope characteristics in the time and frequency domains. Coulson considered some practical details but did not include integer estimation. Multiple repeated training signals were considered by Minn et al. in H. Minn, V. K. Bhargava, and K. B. Letaief, "Robust Timing and Frequency Synchronization for OFDM Systems," IEEE Transactions on Wireless Communications, Vol. 2, No. 4, pp. 822-839, July 2003, but the training structure was not optimized. Shi and Serpedin showed that the training structure had some optimality from the perspective of frame synchronization (K. Shi and E. Serpedin, "Coarse Frame and Carrier Synchronization for OFDM Systems: New Metrics and Comparison," IEEE Transactions on Wireless Communications, Vol. 3, No. 4, pp. 1271-1284, July 2004). In one embodiment of the present invention, the Shi and Serpedin approach is used to provide frame synchronization and fractional frequency offset estimation.

[0296] Many approaches in the literature focus on frame synchronization and fractional frequency offset estimation. Integer offset correction is resolved using additional training symbols, as in T.M. Schmidl and D.C. Cox, "Robust Frequency and Timing Synchronization for OFDM," IEEE Transactions on Communications, Vol. 45, No. 12, pp. 1613-1621, December 1997. For example, Morrelli et al. derived an improved version of T.M. Schmidl and D.C. Cox, "Robust Frequency and Timing Synchronization for OFDM," IEEE Transactions on Communications, Vol. 45, No. 12, pp. 1613-1621, December 1997, in M. Morelli, A.D. Andrea, and U. Mengali, "Resolving Frequency Ambiguity in OFDM Systems," IEEE Communications Letters, Vol. 4, No. 4, pp. 134-136, April 2000. An alternative approach using a different preamble structure has been proposed by Morelli and Mengali (M. Morelli and U. Mengali, "An Improved Frequency Offset Estimator for OFDM Applications," IEEE Communications Letters, Vol. 3, No. 3, pp. 75-77, March 1999). This approach uses correlation between M repeated identical training symbols to extend the range of the fractional frequency offset estimator by a factor of M. This is the best linear unbiased estimator and can tolerate large offsets (with proper design), but it does not provide good timing synchronization.

[0297] System Description One embodiment of the present invention uses channel state information based precoding to cancel frequency and phase offsets in DIDO systems. For a description of this embodiment, see Figure 11 and the related previous discussion.

[0298] In one embodiment of the present invention, each user uses a receiver equipped with a frequency offset estimator / corrector. As shown in Figure 45, in one embodiment of the present invention, a system including a receiver includes a plurality of RF transmitter units 4508, a corresponding plurality of A / D units 4510, a receiver equipped with a frequency offset estimator / corrector 4512, and a DIDO feedback generator unit 4506.

[0299] The RF unit 4508 receives the signal transmitted from the DIDO transmitter unit, downconverts the signal to baseband, and provides the downconverted signal to the A / D unit 4510. The A / D unit 4510 then converts the signal from analog to digital and sends it to the frequency offset estimator / corrector unit 4512. The frequency offset estimator / corrector unit 4512 estimates and corrects for frequency offset as described herein and then sends the correction signal to the OFDM unit 4513. The OFDM unit 4513 removes the cyclic prefix and performs a fast Fourier transform (FFT) to bring the signal into the frequency domain. During the training period, the OFDM unit 4513 sends its output to the channel estimator 4504, which calculates channel estimates in the frequency domain. Alternatively, the channel estimates can be calculated in the time domain. During the data period, the OFDM unit 4513 sends its output to the receiver unit 2002, which demodulates / decodes the signal to obtain data 2014. The channel estimator 4504 sends the channel estimates to a DIDO feedback generator unit 4506 which can quantize the channel estimates and send them to the transmitter over a feedback control channel as shown.

[0300] Description of an embodiment of the algorithm for the DIDO2x2 scenario An embodiment of an algorithm for frequency / phase offset correction in a DIDO system is described below. DIDO system models are first described with and without frequency / phase offsets. For simplicity, a specific example of a DIDO 2x2 system is shown. However, the underlying principles of the present invention can be implemented on higher-order DIDO systems.

[0301] DIDO system model without frequency and phase offsets The received signal for DIDO2x2 can be written for the first user as follows:

[0302]

number

[0303] For the second user:

[0304]

number

[0305] where t is the discrete time index and h mn and w mn are the channel weight and DIDO precoding weight between the m-th user and the n-th transmit antenna, respectively, and x m is the transmitted signal to user m. mn and w mn Note that t is not a function of t because the channel is assumed to be constant over the period between training and data transmission.

[0306] In the presence of frequency and phase offsets, the received signal can be expressed as:

[0307]

number

[0308] and

[0309]

number

[0310] where T s is the symbol period, and ω Tn= 2Πf for the nth transmitting antenna Tn , ω Um = 2Πf of the mth transmitted symbol Um , f Tn and f Um are the actual carrier frequencies (subject to offset) of the nth transmit antenna and the mth user, respectively. mn is the channel h mn Figure 46 shows the DIDO2x2 system model.

[0311] For the time being, the following definitions will be used:

[0312]

number

[0313] This indicates the frequency offset between the mth user and the nth transmit antenna.

[0314] Description of one embodiment of the present invention A method according to one embodiment of the present invention is illustrated in Figure 47. The method comprises the following general stages (including substages as shown): a training period for frequency offset estimation 4701, a training period for channel estimation 4702, and data transmission via DIDO precoding with correction 4703. These stages are described in more detail below.

[0315] (a) Training period for frequency offset estimation (4701) During the first training period, the base station sends one or more training sequences from each transmit antenna to one of the users (4701). As described herein, a "user" is a wireless client device. For the DIDO2x2 case, the signal received by the mth user is denoted by:

[0316]

number

[0317] where p1 and p2 are the training sequences transmitted from the first and second antennas, respectively.

[0318] The mth user can use any form of frequency offset estimator (i.e., convolution with the training sequence) to estimate the offset Δω m1 and Δω m2 From these values, the user then calculates the frequency offset between the two transmit antennas as follows:

[0319]

number

[0320] Finally, the value in (7) is fed back to base station (4701b).

[0321] p1 and p2 in (6) are the values ​​that the user can use to calculate Δω m1 and Δω m2 Note that the carrier frequency offset estimation is designed to be orthogonal so that m can be estimated. Alternatively, in one embodiment, the same training sequence is used over two consecutive time slots, from which the users estimate their offset. Furthermore, to improve the offset estimation in (7), the same calculations described above can be performed for all users in the DIDO system (not just for the mth user), and the final estimate can be an average (weighted) of the values ​​obtained from all users. However, this solution requires more computation time and feedback. Finally, updating the carrier frequency offset estimation is only necessary if the frequency offset varies over time. Therefore, based on the stability of the clock at the transmitter, this stage 4701 of the algorithm can be performed on a long-term basis (i.e., not for every data transmission), thus reducing the feedback overhead.

[0322] (b) Training period for channel estimation (4702) During the training period, the base station first obtains frequency offset feedback of the value of (7) from the mth user or from multiple users. The value in (7) is used to pre-correct the frequency offset at the transmitting side. Then, the base station sends training data to all users for channel estimation (4702a).

[0323] For a DIDO 2x2 system, the signal received at the first user is given by:

[0324]

number

[0325] For the second user, it looks like this:

[0326]

number

[0327] where:

[0328]

number

[0329] and Δt is a random or known delay between the first and second transmissions of the base station, and p1 and p2 are training sequences transmitted from the first and second antennas for frequency offset and channel estimation, respectively.

[0330] Note that pre-correction is only applied to the second antenna in this embodiment. Expanding (8) we get:

[0331]

number

[0332] Similarly, for the second user:

[0333]

number

[0334] Here, the following is true:

[0335]

number

[0336] At the receiving end, the user corrects the residual frequency offset using the training sequences p1 and p2. Then, the user estimates the vector channel through training (4702b).

[0337]

number

[0338] (12) or channel state information (CSI) of these channels is fed back to the base station (4702b), which calculates the DIDO precoder as described in the following subsection.

[0339] (c) DIDO precoding with pre-compensation (4703) The base station receives the channel state information (CSI) in (12) from the user and calculates the precoding weights through block diagonalization (BD) as follows (4703a):

[0340]

number

[0341] where the vector h1 is defined in (12) and wm =[w m1、 w m2 ]. The invention presented here can be applied to any other DIDO precoding method besides BD. The base station also precompensates for frequency offset by using the estimate in (7) and corrects for phase offset by estimating the delay (Δt) between the second training transmission and the current transmission (4703a). Finally, the base station sends data to the user through the DIDO precoder (4703b).

[0342] After this transmit processing, the signal received at user 1 is given by:

[0343]

number

[0344]

number

[0345]

number

[0346] Here, the following is true:

[0347]

number

[0348] Using property (13), we obtain:

[0349]

number

[0350] Similarly, for user 2, we get:

[0351]

number

[0352]

number

[0353] Expanding (16) gives the following:

[0354]

number

[0355] Here, the following is true:

[0356]

number

[0357] Finally, the user computes (4703c) the residual frequency offset and channel estimates for demodulating data streams x1[t] and x2[t].

[0358] Generalization to DIDONxM In this section, the above techniques are generalized to a DIDO system with N transmit antennas and M users.

[0359] i. Training period for frequency offset estimation During the first training period, the signal received by the mth user as a result of the training sequences sent from the N antennas is given by:

[0360]

number

[0361] where p n is the training sequence transmitted from the nth antenna.

[0362] Offset Δω mn , ∀n=1...N, the mth user calculates the frequency offset between the first and nth transmit antennas as follows:

[0363]

number

[0364] Finally, the value in (19) is fed back to the base station.

[0365] ii. Training Period for Channel Estimation , During the training period, the base station first obtains frequency offset feedback with values ​​from the m-th user or from multiple users. The value in (19) is used to pre-correct the frequency offset at the transmitting side. Then, the base station sends training data to all users for channel estimation. For a DIDO NxM system, the signal received by the m-th user is given by:

[0366]

number

[0367] where:

[0368]

number

[0369] and Δt is a random or known delay between the first and second transmissions of the base station. n is the training sequence transmitted from the nth antenna for frequency offset and channel estimation.

[0370] At the receiver, the user receives the training sequence p nThen, each user m is trained to obtain the vector channel:

[0371]

number

[0372] and feeds it back to the base station, which computes the DIDO precoder as described in the following subsection.

[0373] iii. DIDO Precoding with Pre-Compensation The base station receives the channel state information (CSI) in (12) from the user and calculates the precoding weights through block diagonalization (BD) as follows:

[0374]

number

[0375]

number

[0376]

number

[0377] Here, the vector h m is defined in (21), and w m =[w m1 , w m2... , w mN ]. The base station also pre-corrects for frequency offset by using the estimate in (19) and corrects for phase offset by estimating the delay (Δt) between the second training transmission and the current transmission. Finally, the base station sends data to the user through the DIDO precoder.

[0378] After this transmission process, the signal received by user i is given by:

[0379]

number

[0380] Here, the following is true:

[0381]

number

[0382] Using property (22), we obtain the following:

[0383]

number

[0384] Finally, the user computes the residual frequency offset and channel estimate to demodulate the data stream x1[t].

[0385] result Figure 48 shows the SER results of the DIDO2x2 system with and without frequency offset. It can be seen that the proposed method can completely cancel the frequency / phase offset and obtain the same SER as the system without offset.

[0386] Next, we evaluate the sensitivity of the proposed compensation method to frequency offset estimation errors and / or phase variations of the offset in time. Therefore, we rewrite (14) as follows:

[0387]

number

[0388] where ∈ denotes the estimation error and / or variation of the frequency offset between training and data transmission. Note that the effect of ∈ is to destroy the orthogonality performance in (13) so that the interference terms in (14) and (16) are not fully pre-cancelled at the transmitter. As a result, the SER performance deteriorates when increasing the value of ∈.

[0389] Figure 48 shows the SER performance of the frequency offset compensation method for different values ​​of ∈. In these results, T s = 0.3 ms (i.e., a signal with a bandwidth of 3 kHz). For ∈ = 0.001 Hz (or less), the SER performance is observed to be similar to the offset case.

[0390] f. Description of an embodiment of an algorithm for time and frequency offset estimation Additional embodiments for implementing time and frequency offset estimation (4701b in Figure 47) are described below. The transmit signal structures considered are presented in H. Minn, V. K. Bhargava, and K. B. Letaief, "Robust Timing and Frequency Synchronization for OFDM Systems," IEEE Transactions on Wireless Communications, Vol. 2, No. 4, pp. 822-839, July 2003, and considered in more detail in K. Shi and E. Serpedin, "Coarse Frame and Carrier Synchronization for OFDM Systems: New Metrics and Comparisons," IEEE Transactions on Wireless Communications, Vol. 3, No. 4, pp. 1271-1284, July 2004. Typically, sequences with good correlation performance are used for training. For example, for the system of the present invention, we use Chu sequences derived as described in D. Chu, "Polyphase Codes with Good Periodic Correlation Properties (Concordance)," IEEE Transactions on Information Theory, Vol. 52, No. 4, pp. 531-532, July 1972. These sequences have the interesting property of having perfect circular correlation. cp denotes the length of the cyclic prefix, and N t Let N denote the length of the constituent training sequences. t =M t Here, Mt is the length of the training sequence. With these assumptions, the transmitted symbol sequence of the preamble can be written as:

[0391]

number

[0392] This training signal structure can be extended to other lengths, but note that the block structure is repeated. For example, to use 16 training signals, consider the following structure:

[0393]

number

[0394] Using this structure, N t =4M t By doing so, all the algorithms described can be used without modification. In effect, the training sequence is repeated. This is particularly useful when a suitable training signal is not available.

[0395] After matched filtering and downsampling to the symbol rate, consider the received signal:

[0396]

number

[0397] where ∈ is the unknown discrete-time frequency offset, Δ is the unknown frame offset, h[l] is the unknown discrete-time channel coefficient, and v[n] is additive noise. To illustrate the key ideas in the following sections, we ignore the presence of additive noise.

[0398] i. Coarse Frame Synchronization The goal of coarse frame synchronization is to solve for the unknown frame offset Δ. The following definitions are defined:

[0399]

number

[0400] The proposed coarse frame synchronization algorithm is inspired by the algorithm in K. Shi and E. Serpedin, "Coarse Frame and Carrier Synchronization for OFDM Systems: New Metrics and Comparisons," IEEE Transactions on Wireless Communications, Vol. 3, No. 4, pp. 1271-1284, July 2004, which is derived from a maximum likelihood criterion. Method 1 - Improved Coarse Frame Synchronization : The coarse frame synchronization estimator solves the following optimization:

[0401]

number

[0402] The correction signal is defined as follows:

[0403]

number

[0404] An additional correction term is used to compensate for the small initial taps in the channel and can be adjusted based on the application. This extra delay will hereafter be included in the channel.

[0405] ii. Fractional frequency offset correction Fractional frequency offset correction occurs after the coarse frame synchronization block. Method 2 - Improved fractional frequency offset correction : The fractional frequency offset is the solution to:

[0406]

number

[0407] This is known as fractional frequency offset because the algorithm can only correct for offsets below this value.

[0408]

number

[0409] This problem will be solved in the next section. The fine frequency offset correction signal is defined as follows:

[0410]

number

[0411] It should be noted that Methods 1 and 2 are improvements of K. Shi and E. Serpedin, "Coarse Frame and Carrier Synchronization for OFDM Systems: New Metrics and Comparisons," IEEE Transactions on Wireless Communications, Vol. 3, No. 4, pp. 1271-1284, July 2004, which perform better in frequency-selective channels. One particular innovation here is the use of r and

[0412]

number

[0413] The use of both.

[0414]

number

[0415] The use of improves on the previous estimator by ignoring samples that are likely to be corrupted by intersymbol interference.

[0416] iii. Integer frequency offset correction To compensate for integer frequency offsets, it is necessary to write an equivalent system model for the received signal after fine frequency offset compensation. After absorbing the remaining timing error into the channel, the received signal in the noiseless case is t -1 has the following structure:

[0417]

number

[0418] The integer frequency offset is k, while the unknown equivalent channel is g[l]. Method 3 - Improved Integer Frequency Offset Correction : The integer frequency offset is the solution to:

[0419]

number

[0420] Here, the following is true:

[0421]

number

[0422]

number

[0423]

number

[0424]

number

[0425] This gives an estimate of the total frequency offset as follows:

[0426]

number

[0427] In practice, the complexity of Method 3 is quite high. To reduce the complexity, the following observations can be made. First, the product:

[0428]

number

[0429] can be pre-computed. Unfortunately, this still requires a fairly large matrix multiplication. An alternative is to use

[0430]

number

[0431] The key to success is to take advantage of the observation that Method 4 - Low Complexity Improved Integer Frequency Offset Correction: The low complexity integer frequency offset estimator solves:

[0432]

number

[0433] iv.Results In this section, we compare the performance of different proposed estimators.

[0434] First, Figure 50 compares the amount of overhead required for each method. Note that both new methods reduce the overhead required by a factor of 10 to 20. Monte Carlo experiments were performed to compare the performance of the different estimators. The configuration considered is the inventors' typical NVIS transmit waveform, consisting of a linear modulation with a 3 kHz passband and a symbol rate of 3K symbols / second, corresponding to a raised cosine pulse shape. For each Monte Carlo run, the frequency offset is [-f max , f max ] is generated from a uniform distribution of

[0435] f max Simulations with a small frequency offset of N = 2 Hz and no integer offset correction are shown in Figure 51. From this performance comparison, t / M t Although the performance with N = 1 is slightly worse than the original estimator, we can still see that the overhead is significantly reduced. t / M t The performance with = 4 is much better, by almost 10 dB. All curves have a knee at low SNR points due to errors in integer offset estimation. Small errors in the integer offset can cause large frequency errors, and therefore large mean square errors. The integer offset correction can be turned off at small offsets to improve performance.

[0436] In the presence of multipath channels, the performance of the frequency offset estimator usually deteriorates. However, with the integer offset estimator turned off, very good performance is shown in Figure 52. Therefore, in multipath channels, it is even more important to implement a robust coarse correction and then an improved fine correction algorithm. N t / M t Note that the offset performance with =4 is much better in the multipath case.

[0437] Embodiments of the present invention may include various steps as described above. These steps may be implemented with machine-executable instructions that cause a general-purpose or special-purpose processor to perform certain steps. For example, they may be implemented as software running on general-purpose or special-purpose processors in various components within the base station / AP and the client devices described above. Various well-known personal computer components, such as computer memory, hard drives, input devices, etc., have not been shown to avoid obscuring relevant aspects of the present invention.

[0438] Alternatively, in one embodiment, the various functional modules and associated steps illustrated herein may be implemented by specific hardware components including hardwired logic that implements the steps, such as an application specific integrated circuit (ASIC), or by any combination of programmed computer components and custom hardware components.

[0439] In one embodiment, certain modules, such as the encoding, modulation, and signal processing logic 903 described above, may be implemented on a programmable digital signal processor (DSP) (or a group of DSPs), such as a DSP using Texas Instruments' TMS320x architecture (e.g., TMS320C6000, TMS320C5000, etc.). The DSP in this embodiment may be embedded within an add-on card to a personal computer, such as a PCI card. Of course, a variety of different DSP architectures may be used while still adhering to the underlying principles of the present invention.

[0440] Elements of the present invention may also be provided as a machine-readable medium storing machine-executable instructions. The machine-readable medium may include, but is not limited to, flash memory, optical disks, CD-ROMs, DVD-ROMs, RAM, EPROMs, EEPROMs, magnetic or optical cards, propagation media, or any other form of machine-readable medium suitable for storing electronic instructions. For example, the present invention may be downloaded as a computer program product that may be transferred from a remote computer (e.g., a server) to a requesting computer (e.g., a client) over a communications link (e.g., a modem or network connection) by data signal embodied in a carrier wave or other propagation medium.

[0441] Throughout the foregoing description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the systems and methods of the present invention. However, it will be apparent to one skilled in the art that the systems and methods may be practiced without some of these specific details. Accordingly, the scope and spirit of the present invention should be determined by the terms of the following claims.

[0442] Furthermore, throughout the foregoing description, numerous references have been cited to provide a more complete understanding of the present invention, all of which are incorporated herein by reference.

Claims

1. one or more radio transceiver stations configured to form multiple simultaneous and independent radio links with multiple user equipments of the same frequency band, the radio links forming downlink or uplink channels; the one or more wireless transceiver stations are configured to form at least one of the multiple simultaneous and independent wireless links, with a number of wireless transceiver station antennas equal to at least 10 times the number of antennas of at least one user device configured to communicate over the multiple simultaneous and independent wireless links.

2. 2. The system of claim 1, wherein said wireless links have a particular wavelength, and said antennas of said radio transceiver stations are spaced apart at intervals less than said particular wavelength.

3. The system of claim 1 , wherein one or a subset of the antennas are orthogonally polarized.

4. 2. The system of claim 1, wherein the wireless transceiver station comprises at least ten times as many antennas as the number of antennas configured for communicating with at least one of the plurality of simultaneous and independent wireless links at each user device.

5. 2. The system of claim 1, wherein the wireless transceiver station is configured to communicate with the at least one user device equipped with multiple antennas configured to communicate with the plurality of simultaneous and independent wireless links.

6. 2. The system of claim 1, wherein the wireless transceiver station is configured to communicate with user equipment equipped with between two and four antennas configured to communicate with the at least one of the plurality of simultaneous and independent wireless links.

7. 2. The system of claim 1, wherein the wireless transceiver station is configured to communicate with the at least one user device equipped with a single antenna configured to communicate with the plurality of simultaneous and independent wireless links.

8. one or more radio transceiver stations configured to form multiple simultaneous and independent radio links with multiple user equipments of the same frequency band, the radio links forming downlink or uplink channels; the one or more wireless transceiver stations are configured to form the plurality of simultaneous and independent wireless links, the number of wireless transceiver station antennas being equal to at least ten times the number of antennas of at least one user device configured to communicate over the at least one simultaneous and independent wireless link.

9. 10. The system of claim 8, wherein the wireless transceiver station is configured to communicate with user equipment equipped with multiple antennas configured to communicate over at least one of the simultaneous and independent wireless links.

10. 10. The system of claim 8, wherein the wireless transceiver station is configured to communicate with user devices equipped with between two and four antennas configured to communicate over at least one of the simultaneous and independent wireless links.

11. 9. The system of claim 8, wherein the wireless transceiver station is configured to communicate with a user device equipped with a single antenna configured to communicate over at least one of the simultaneous and independent wireless links.

12. 9. The system of claim 8, wherein the antenna on the radio transceiver station is a directional antenna.

13. 10. The system of claim 8, wherein the antenna is a phased array implementing beam steering.

14. 10. The system of claim 8, wherein the wireless transceiver stations are configured to beamform to increase link reliability or channel capacity, beamforming including forming multiple simultaneous beams on the fly.

15. 9. The system of claim 8, wherein the radio transceiver station comprises at least ten times as many antennas as each user device.

16. one or more radio transceiver stations configured to form multiple simultaneous and independent radio links with multiple user equipments of the same frequency band, the radio links forming downlink or uplink channels; the one or more wireless transceiver stations are configured to form the plurality of simultaneous and independent wireless links, the plurality of wireless transceiver station antennas having a number equal to at least ten times the number of the at least one plurality of simultaneous and independent wireless link antennas of at least one user device.

17. 17. The system of claim 16, wherein the transceiver station is configured to utilize channel interactions between the multiple antennas on the transceiver station and one or more antennas of the multiple user devices to estimate uplink channel state information (CSI) and use this to derive downlink channel state information (CSI).

18. 17. The system of claim 16, wherein the transceiver station is configured to estimate channel state information (CSI) between the plurality of antennas on the transceiver station and one or more antennas of the plurality of user devices by using pilot tones, channel estimation, or training signals.

19. 17. The system of claim 16, wherein the transceiver station is configured to estimate channel state information (CSI) between the plurality of antennas on the transceiver station and one or more antennas of the plurality of user devices via orthogonal training signals.

20. 17. The system of claim 16, wherein the transceiver station is configured to store channel state information (CSI) estimates between the plurality of antennas on the transceiver station and one or more antennas of the plurality of user devices in a channel characterization matrix.

21. 17. The system of claim 16, wherein the transceiver station is configured to store channel state information (CSI) estimates between the plurality of antennas on the transceiver station and one or more antennas of the plurality of user devices in the channel characterization matrix, the channel characterization matrix including the phase and amplitude of the channel.

22. 17. The system of claim 16, wherein the transceiver station is configured to transmit training signals to user equipment.

23. 17. The system of claim 16, wherein the transceiver station is configured to transmit synchronized training signals towards user equipment.

24. 17. The system of claim 16, wherein the transceiver station is configured to continuously or periodically update channel state information (CSI) estimates between the plurality of antennas on the transceiver station and one or more antennas of the plurality of user devices.

25. 17. The system of claim 16, wherein the transceiver station is configured to update channel state information (CSI) estimates between the plurality of antennas on the transceiver station and one or more antennas of the plurality of user devices when the user devices move from one location to another.

26. 17. The system of claim 16, wherein the transceiver station is configured to form the data link through precoding based on channel state information (CSI) estimates between the plurality of antennas on the transceiver station and one or more antennas of the plurality of user devices.

27. 17. The system of claim 16, wherein the transceiver station is configured to divide the user equipment into a plurality of groups.

28. 17. The system of claim 16, wherein the transceiver station is configured to divide the user equipment into a plurality of groups and to allocate different groups of users at times by cycling between the groups.

29. 17. The system of claim 16, wherein the transceiver station is configured to divide user equipment into groups and allocate the same or different levels of bandwidth to the user equipment.

30. 17. The system of claim 16, wherein the transceiver station is configured to divide the user devices into groups and assign the groups over time based on relative proximity between the user devices.

Images