Next Generation Mobile Communication Channel State Information (CSI) Acquisition for mmWave MIMO Systems Mi Michael hael Liao ao Advisor : Andy Wu Graduate Institute of Electronics Engineering National Taiwan University Taipei, Taiwan April 7, 2015
Outline Recap: Millimeter Wave & Hybrid-Beamforming Why Do We Need “Channel State Information” ? CSI Acquisition in Hybrid Architecture Existing Solutions for CSI acquisition Summary Future work Reference 1 ACCESS
Recap: Why mmWave for Cellular ? 8GHz 11GHz 20GHz Cellular systems live with a little microwave spectrum 600MHz total (best case) in the U.S. Spectrum efficiency is king (MIMO, MU-MIMO) Huge amount of spectrum available in mmWave bands Unlicensed and free spectrum for links of very high data rates mmWave already used in LAN and PAN. mmWave links are used for backhaul in cellular networks 2 ACCESS
Recap: Big Challenges from mmWave MIMO (1/3) 30 GHz shows additional about 20 20 dB dB loss compared to 3 GHz. Issues from large pathloss in mmWave – Links with unreasonable signal-to-noise ratio (SNR) Rank-deficient MIMO due to reduced spatial dimension Measurements of Propagation Loss [1] Tx X NLOS X LOS Rx NLOS mmWave Transmission Microwave Transmission 3 ACCESS
Recap: Big Challenges from mmWave MIMO (2/3) Pathloss in mmWave can be compensated through large array antenna and directional transmissions. Huge be beam amform rming gai gain with compact-sized array antenna Meanwhile multiplexing gain still can be achieved by conventional MIMO techniques (e.g. P2P-MIMO, MU-MIMP) Measurements of Propagation Loss Wireless Multipath Propagation Tx Distance (m) Rx Channel Tx Array Antenna Rx Array Antenna 4 ACCESS
Recap: Big Challenges from mmWave MIMO (3/3) (1) Large antenna array requires multiple RF/analog chains. Analog-to-digital/digital-to-analog converters Required power amplifiers are proportional to antennas (2) Higher dimension information of MIMO channel acquisition Estimation of channel state information (CSI) Number of Total Tx Antennas : M Training & controlling resource overhead Number of Total Rx Antennas : N Number of Independent Data Stream : K (2) (1) 5 ACCESS
Recap: Hybrid-Beamforming Technique Traditional Beamforming is done at BB Requiring one RF chain per transmitting antenna. Hybrid Beamforming relies on RF precoding to reduce the number of RF chains arg min Two-staged structure: ( F , F ) F F F BB RF opt RF BB F , F F RF BB 6 ACCESS
Why Do We Need “Channel State Information” ? In wireless communications, channel state information (CSI) refers to known channel properties of a communication link. CSI needs to be estimated at the receiver and usually quantized and feedback to the transmitter. V1 must be acquired from performing SVD on H. 7 ACCESS
CSI Acquisition in Hybrid Architecture (1/4) mmWave MIMO with Hybrid Architecture H Due to the hybrid architecture, CSI cannot be acquired by conventional methods directly. Account for practical assumptions on the hybrid architecture: RF chains is limited, i.e., less than the number of antennas Phase shifters have constant modulus and quantized phases 8 ACCESS
CSI Acquisition in Hybrid Architecture (2/4) RF beamformers greatly reduces the complexity & cost Requires fewer RF chains However, it cannot “see” the whole channel instantaneously Instead, we have to “scan” the channel with directional beams Controlled by analog switches or phase shifters 9 ACCESS
CSI Acquisition in Hybrid Architecture (3/4) 10 ACCESS
CSI Acquisition in Hybrid Architecture (4/4) Angular Domain Transform [7] CSI acquisition can be separated in two parts: (1) Beamformer Training: Find out the active spatial channel path. (2) Channel Estimation: Estimation each entry (path gain) in the H a 11 ACCESS
Existing Solutions for CSI acquisition (1/2) Conventional beam scanning protocol Exhaustive search over a large space given the large # of antennas Scanned requires fine control and leads to slow adaptation Trading overhead – O(N N ) BS MS Exhaustive Search 12 ACCESS
Existing Solutions for CSI acquisition (2/2) Multi-resolution codebook for beamforming vector [8]-[11] Different granularity of beam patterns from codebook and realize progressive searching Greatly reduces the training overhead for large antenna array Quasi-Omni Sector Single Path 8x8 System Beam Hi-res Beam Exhaustive Search = 8x8 = 64 Progressive Search = 3x(2x2) = 12 Multipath 8x8 System by Progressive Search = 12 * (# of paths) 13 ACCESS
Issues of Existing Solutions 1. However, multi-resolution codebooks is subject to practical limitations in our applications. Existence of quantized phase shifters makes the design of non-overlapping beam patterns difficult The Codebook design of this work relies mostly on 2D array. Hybrid analog-digital training codebook is required 2. Existing solutions are not designed for MIMO CSI acquisition How to estimate the parameters of path gain ? Conventional progressive search is not designed for multipath systems H H = U H U r a t H | | | | | | a ( ) a ( ) a ( ) a ( ) a ( ) a ( ) MS 1 MS 2 MS N BS 1 BS 2 BS N MS BS | | | | | | 14 ACCESS
Summary & Future Work Introduction to Channel State Information. For CSI acquisition in the hybrid precoding transceiver, we prefer to use: Channel estimation from angular domain . Hybrid analog-digital training codebook CSI acquisition can be separated in two parts: Step I: Beamform Training Beam Alignment & Beamform Tracking Step II: Channel Estimation 15 ACCESS
What is “ Beamform Tracking”? Rapid proliferation of smart Beam Tracking in BS phones and tablets. Beam alignment can be easily destroyed by even small variation. Z4 Re-estimating the channel Z3 Z2 causes high overhead. Z1 Track the channel variations during user mobility with low overhead. MS1 MS1 16 ACCESS
Reference [1] C. Yungwhoon, F. Aryanfar, "Millimeter-wave beamforming as an enabling technology for 5G cellular communications: theoretical feasibility and prototype results," IEEE Comm. Mag., IEEE , vol.52, no.2, pp.106-113, Feb. 2014 [2] P. Sudarshan, N. B. Mehta, A. D. Molisch and J. Zhang, “Channel statistics-based RF pre-processing with antenna selection,” IEEE Trans. on Wireless Communications, vol. 5, no. 12, pp. 3501-3511, Dec. 2006. [3] Y.-S. Yin, L.-K. Chiu and S.-H. Wu, “A space-time precoded hybrid beamforming architecture for broadband transmissions in 60GHz radio,” in Proc. 2011 PIMRC. [4] S.-S. Wu, L.-K. Chiu, K.-Y. Lin, and T.-H. Chang, “Robust hybrid beamforming with phased antenna arrays for downlink SDMA in indoor 60 GHz channels,” IEEE Trans. on Wireless Communications, vol. 12, no. 9, pp.4542-4557, Sept. 2013. [5] J. Guo, X. Huang and V. Dyadyuk, “A hybrid adaptive antenna array for long-range mm-Wave communication,” IEEE Antennas and Propagation Magazine, vol. 54, no. 2, pp. 271-282, April 2012 [6] J. Nsenga, A. Bourdoux, and F. Horlin, “Mixed analog/digital beamforming for 60GHz MIMO frequency selective channels,” in Proc. 2010 IEEE International Conf. Commun (ICC). [7] D. Tse and P. Viswanath, Fundamentals of wireless communication. Cambridge Univ Pr, 2005. [8] J. Wang, Z. Lan, C. Pyo, T. Baykas, C. Sum, M. Rahman, J. Gao, R. Funada, F. Kojima, H. Harada et al., “Beam codebook based beamforming protocol for multi-Gbps millimeter-wave WPAN systems,” IEEE Journal on Selected Areas in Communications, vol. 27, no. 8, pp. 1390 – 1399, 2009. [9] L. Chen, Y. Yang, X. Chen, and W. Wang, “Multi -stage beamforming codebook for 60GHz WPAN,” in Proc. of 6 th International ICST Conference on Communications and Networking in China, China, 2011, pp. 361 – 365. [10] S. Hur, T. Kim, D. Love, J. Krogmeier, T. Thomas, and A. Ghosh, “Millimeter wave beamforming for wireless backhaul and access in small cell networks,” IEEE Transactions on Communications, vol. 61, no. 10, pp. 4391 – 4403, 2013. [11] Y. M. Tsang and A. S. Poon, “Successive AoA estimation: Revealing the second path for 60 GHz communication system,” 49th Annual Allerton Conference on Communication, Control, and Computing (Allerton), pp. 508 – 515, 2011. [12] A. Alkhateeb, O. E. Ayach, and G. Leus, “Channel Estimation and Hybrid Precoding for Millimeter Wave Cellular Systems,” submitted to IEEE Journal of Selected Topics in Signal Processing, on Jan 2014 17 ACCESS
Recommend
More recommend
