#### DMCA

## 1 Antenna Grouping based Feedback Reduction for FDD-based Massive MIMO Systems

### Citations

588 |
On the achievable throughput of a multiantenna Gaussian broadcast channel
- Caire, Shamai
- 2003
(Show Context)
Citation Context ... + P K |h H k wk| 2 1 + PK ∑K j=1,j 6=k |h H k wj | 2 ) (3) and the corresponding sum rate becomes Rsum = K∑ k=1 Rk. In generating the beamforming vectors, we consider zero-forcing beamforming (ZFBF) =-=[16]-=-–[18] where the right pseudo inverse Wzf = H H ( HHH )−1 of the channel matrix H is applied to the message vector s[n] to remove the inter-user interference.1 In order to satisfy the transmit power co... |

329 | Grassmannian beamforming for multiple-input multiple-output wireless systems
- Love, Heath, et al.
- 2003
(Show Context)
Citation Context ...of nearby antenna elements may be a desirable option in practice (see the example in Fig. 4). Alternatively, one can consider Grassmannian subspace packing in the design of the antenna group patterns =-=[21]-=-. The main goal of Grassmannian subspace packing is, when the subspace distance metric and the number of feedback bits B are provided, to find a set of 2B subspaces in2 G(Nt,m) that maximizes the mini... |

308 | On the optimality of multiantenna broadcast scheduling using zero-forcing beamforming,” - Yoo, Goldsmith - 2006 |

274 |
Noncooperative cellular wireless with unlimited numbers of base station antennas
- Marzetta
- 2010
(Show Context)
Citation Context ...-scale transmit antenna array, often called massive MIMO, have been of great interest in recent years because of their potential to dramatically improve spectral efficiency of future wireless systems =-=[2]-=-, [3]. By employing tens or hundreds of antennas at the basestation, massive MIMO systems can control intra-cell interference and thermal noise by simply using linear precoding in the downlink and rec... |

205 | An overview of limited feedback in wireless communication systems
- Love, Heath, et al.
- 2008
(Show Context)
Citation Context ...re is on-going debate regarding the pros and cons of time division duplexing (TDD) and frequency division duplexing (FDD). In obtaining the CSI, FDD requires the CSI to be fed back through the uplink =-=[6]-=- while no such procedure is required for TDD systems owing to channel reciprocity [7]. In fact, under the assumption that RF chains are properly calibrated [8], the CSI of the downlink can be estimate... |

155 | MIMO broadcast channels with finite rate feedback,”
- Jindal
- 2006
(Show Context)
Citation Context ...orrelation among antenna elements and the random vector quantization (RVQ) is used, the number of feedback bits per user should be scaled with the number of transmit antennas and SNR (in decibels) as =-=[13]-=-, [19] Buser = (Nt − 1) log2 P ≈ Nt − 1 3 PdB (8) to maintain a constant gap in terms of the sum rate from the system with perfect CSI. Clearly, when the number of transmit antennas increases, the fee... |

139 |
On the asymptotic eigenvalue distribution of Toeplitz matrices
- Gray
- 1972
(Show Context)
Citation Context ...cient, and α is the magnitude of correlation coefficient, and θ is the phase of the user. When the number of transmit antennas Nt is large, (non-ordered) singular value µi of Rt becomes approximately =-=[25]-=- µi ≈ Nt−1∑ k=−(Nt−1) ρ|k|ej 2piik Nt ≈ 1− ρ2 1 + ρ2 − 2ρ cos(2piiNt ) , i = 1, . . . , Nt. (27) Using the first and second largest singular values (i.e., µNt , µNt−1) of (27), we have σ2 σ1 ≈ 1 + ρ2 ... |

119 | Multi-antenna downlink channels with limited feedback and user selection
- Yoo, Jindal, et al.
- 2007
(Show Context)
Citation Context ...2 ] = E [ ‖h‖2 ( 1− |h̄H h̃r| 2 )] (16) where h̃r is the expanded version of the quantized vector ĥr (see (12)). In order to evaluate the distortion D, we use the quantization cell upper bound (QUB) =-=[23]-=-. As mentioned, since a codebook designed for the i.i.d channels is not the right choice for correlated channels, we employ a channel statistic-based codebook obtained by applying 6Antenna groupspatte... |

116 | Channel capacity of MIMO architecture using the exponential correlation matrix
- Loyka
- 2001
(Show Context)
Citation Context ...the feedback allocation, we use the notation B = (Bq, Bp, Br) in the sequel. As a channel model, we consider the exponential correlation model in (26), onedimensional uniform linear array (ULA) model =-=[26]-=-, and two-dimensional uniform planar array (UPA) model [27]. As a channel vector quantizer, we use RVQ and noncoherent trellis coded quantization (NTCQ). While RVQ is a well-known and standard channel... |

109 | Massive MIMO in the UL/DL of cellular networks: How many antennas do we need
- Hoydis, Brink, et al.
- 2013
(Show Context)
Citation Context ...e transmit antenna array, often called massive MIMO, have been of great interest in recent years because of their potential to dramatically improve spectral efficiency of future wireless systems [2], =-=[3]-=-. By employing tens or hundreds of antennas at the basestation, massive MIMO systems can control intra-cell interference and thermal noise by simply using linear precoding in the downlink and receive ... |

108 |
On the performance of random vector quantization limited feedback beamforming in a MISO system,” Wireless Communications,
- Au-Yeung, Love
- 2007
(Show Context)
Citation Context ...from FDD-based massive MIMO techniques. One well-known problem of FDD system is that the amount of CSI feedback must scale linearly with the number of antennas to control the quantization error [12]– =-=[14]-=-. Therefore, it is not hard to convince oneself that the overhead of CSI feedback is a serious concern when the number of transmit antennas is large. Needless to say, a technique that efficiently redu... |

76 | Pilot contamination and precoding in multi-cell tdd systems
- Jose, Ashikhmin, et al.
- 2011
(Show Context)
Citation Context ...d frequency division duplexing (FDD). In obtaining the CSI, FDD requires the CSI to be fed back through the uplink [6] while no such procedure is required for TDD systems owing to channel reciprocity =-=[7]-=-. In fact, under the assumption that RF chains are properly calibrated [8], the CSI of the downlink can be estimated using the pilot signal in the uplink so that the CSI feedback information from the ... |

64 | Multiple antenna broadcast channels with shape feedback and limited feedback,”
- Ding, Love, et al.
- 2007
(Show Context)
Citation Context ...tion among antenna elements and the random vector quantization (RVQ) is used, the number of feedback bits per user should be scaled with the number of transmit antennas and SNR (in decibels) as [13], =-=[19]-=- Buser = (Nt − 1) log2 P ≈ Nt − 1 3 PdB (8) to maintain a constant gap in terms of the sum rate from the system with perfect CSI. Clearly, when the number of transmit antennas increases, the feedback ... |

47 | Argos: Practical many-antenna base stations.
- Shepard, Yu, et al.
- 2012
(Show Context)
Citation Context ... in the uplink so that the CSI feedback information from the user terminal to the basestation is unnecessary. Due to this benefit, most of the massive MIMO works in the literature have focused on TDD =-=[9]-=- (possible exceptions are [10], [11]). However, since FDD dominates current cellular networks and offers many benefits over TDD (e.g., small latency, continuous channel estimation, backward compatibil... |

39 |
Asymptotic performance of MIMO wireless channels with limited feedback,” in
- Santipach, Honig
- 2003
(Show Context)
Citation Context ...ising from FDD-based massive MIMO techniques. One well-known problem of FDD system is that the amount of CSI feedback must scale linearly with the number of antennas to control the quantization error =-=[12]-=-– [14]. Therefore, it is not hard to convince oneself that the overhead of CSI feedback is a serious concern when the number of transmit antennas is large. Needless to say, a technique that efficientl... |

39 |
Joint spatial division and multiplexing – the large-scale array regime,”
- Adhikary, Nam, et al.
- 2013
(Show Context)
Citation Context ...r) in the sequel. As a channel model, we consider the exponential correlation model in (26), onedimensional uniform linear array (ULA) model [26], and two-dimensional uniform planar array (UPA) model =-=[27]-=-. As a channel vector quantizer, we use RVQ and noncoherent trellis coded quantization (NTCQ). While RVQ is a well-known and standard channel vector quantizer in the literatures, the time to choose th... |

38 |
Asymptotic eigenvalue distribution and capacity for MIMO channels under correlated fading
- Martin, Ottersten
- 2004
(Show Context)
Citation Context ...correlation model is used, however, we can observe the relationship between two. For example, if the exponential correlation model is employed, then the transmit correlation matrix Rt is expressed as =-=[24]-=- Rt = 1 ρ · · · ρNt−1 ρH 1 · · · ρNt−2 ... ... . . . ... ρ(Nt−1)H ρ(Nt−2)H · · · 1 (26) where ρ = αejθ is the transmit correlation coefficient, and α is the magnitude of correlation coef... |

29 | Massive MIMO Systems With Non-Ideal Hardware: Energy Efficiency, Estimation, and Capacity Limits,”
- Bjornson, Hoydis, et al.
- 2014
(Show Context)
Citation Context ...he CSI to be fed back through the uplink [6] while no such procedure is required for TDD systems owing to channel reciprocity [7]. In fact, under the assumption that RF chains are properly calibrated =-=[8]-=-, the CSI of the downlink can be estimated using the pilot signal in the uplink so that the CSI feedback information from the user terminal to the basestation is unnecessary. Due to this benefit, most... |

23 |
Limited feedback diversity techniques for correlated channels
- Love, Heath, et al.
- 2006
(Show Context)
Citation Context ...· · , c2B−Bp}. We note that since a codebook designed for i.i.d channels is not a proper choice for correlated channels, we consider a channel statistic-based codebook for channel vector quantization =-=[20]-=- (see Section III.C for details). The codeword ĥ (i) r maximizing the absolute inner product with h (i) r is chosen as ĥ(i)r = argmax c∈C |h̄(i)Hr c| 2, i = 1, · · · , NP (11) where h̄ (i) r = h (i)... |

19 |
Joint spatial division and multiplexing: Realizing massive MIMO gains with limited channel state information
- Nam, Ahn, et al.
- 2012
(Show Context)
Citation Context ...back information from the user terminal to the basestation is unnecessary. Due to this benefit, most of the massive MIMO works in the literature have focused on TDD [9] (possible exceptions are [10], =-=[11]-=-). However, since FDD dominates current cellular networks and offers many benefits over TDD (e.g., small latency, continuous channel estimation, backward compatibility), it is important to identify an... |

14 |
Full-dimension MIMO (FD-MIMO) for next generation cellular technology,”
- Nam
- 2013
(Show Context)
Citation Context ...012R1A2A2A01047510). the transmit power of each terminal inversely proportional to the number of basestation antennas [3]. Presently, standardization activity for massive MIMO has been initiated [4], =-=[5]-=-, and there is on-going debate regarding the pros and cons of time division duplexing (TDD) and frequency division duplexing (FDD). In obtaining the CSI, FDD requires the CSI to be fed back through th... |

11 |
Noncoherent trellis coded quantization: A practical limited feedback technique for massive MIMO systems,”
- Choi, Chance, et al.
- 2013
(Show Context)
Citation Context ...I feedback information from the user terminal to the basestation is unnecessary. Due to this benefit, most of the massive MIMO works in the literature have focused on TDD [9] (possible exceptions are =-=[10]-=-, [11]). However, since FDD dominates current cellular networks and offers many benefits over TDD (e.g., small latency, continuous channel estimation, backward compatibility), it is important to ident... |

11 | Multi-User Diversity vs. Accurate Channel State Information
- Ravindran, Jindal
- 2009
(Show Context)
Citation Context ...t codebook is used, we need to check 2B codewords). Since the number of transmit antennas Nt in our simulations is large (i.e, 32 and 64 antennas), we use the analytical closed-form expression of RVQ =-=[28]-=- instead of simulations. To make sure that the analysis result can be translated into a realizable codebook, we also use the recently proposed low-complexity codebook called NTCQ [10]. Note that the s... |

10 |
Correlated fading in broadcast MIMO channels: Curse or blessing
- Clerckx, Kim, et al.
- 2008
(Show Context)
Citation Context ...rly to the number of transmit antennas so that one can evaluate the performance of NTCQ codebook with reasonable simulation time. B. Simulation Results We first consider the exponential channel model =-=[29]-=- rij = { ρ |j−i| k i ≤ j (ρ |j−i| k ) H i > j (30) where rij is the (i, j)-th element of Rt,k and ρk = αe jθk is a transmit correlation coefficient for the k-th user (α is the magnitude of correlation... |

4 |
Recent Trend of Multiuser MIMO in LTE-Advanced
- Lim, Yoo, et al.
- 2013
(Show Context)
Citation Context ...with user scheduling [31] would be interesting direction to be investigated. Finally, we note that the proposed method can be nicely integrated into the dual codebooks structure in LTE-Advanced [32], =-=[33]-=- by feeding back the pattern index for long-term basis and the codebook index for short-term basis. Since the main target of the massive MIMO system is slowly varying or static channels, dual codebook... |

1 |
MU-MIMO with channel statistics based codebooks in spatially correlated channels
- Clerckx, Kim, et al.
- 2008
(Show Context)
Citation Context ...where Rt,k ∈ CNt×Nt is the transmit correlation matrix of the k-th user and hw,k ∈ CNt is the channel vector whose elements are independent and identically distributed according to hw,k ∼ CN (0, INt) =-=[15]-=-. The matrix-vector form of (1) is expressed as y[n] = Hx[n] + z[n] (2) where H = [h1 h2 . . . hK ] H ∈ CK×Nt is the composite channel matrix, z[n] = [z1[n] z2[n] . . . zK [n]] T ∈ CK is the complex G... |

1 |
A MMSE vector precoding with block diagonalization for multiuser MIMO downlink
- Park, Lee, et al.
- 2012
(Show Context)
Citation Context ...K |h H k wk| 2 1 + PK ∑K j=1,j 6=k |h H k wj | 2 ) (3) and the corresponding sum rate becomes Rsum = K∑ k=1 Rk. In generating the beamforming vectors, we consider zero-forcing beamforming (ZFBF) [16]–=-=[18]-=- where the right pseudo inverse Wzf = H H ( HHH )−1 of the channel matrix H is applied to the message vector s[n] to remove the inter-user interference.1 In order to satisfy the transmit power constra... |

1 |
Correlation matrix diatance, a meaningful measure for evaluation of nonstationary
- Herdin, Czink, et al.
- 2005
(Show Context)
Citation Context ...erns (expansion matrices) maximizing the minimum distance metric between any pair of subspaces. To measure the distance, we use the correlation matrix distance dcorr(A,B) between two matrices A and B =-=[22]-=- dcorr(A,B) = 1− tr(AHB) ‖A‖F‖B‖F . (15) Note that dcorr(A,B) measures the orthogonality between two correlation matrices A and B. When the correlation matrices are equal up to a scaling factor, dcorr... |

1 |
Downlink training techniques for FDD masive MIMO systems: open-loop and closedloop training with memory,” accepted for
- Choi, Love, et al.
(Show Context)
Citation Context ...B algorithm can be used as an effective means to achieve reduction in the feedback information. Also, performance evaluation of the AGB algorithm under channel estimation through closed-loop training =-=[30]-=- and investigation of nonlinear transmitter techniques with user scheduling [31] would be interesting direction to be investigated. Finally, we note that the proposed method can be nicely integrated i... |

1 |
A vector perturbation with user selection for multiuser MIMO downlink
- Lee, Shim
- 2012
(Show Context)
Citation Context ...ck information. Also, performance evaluation of the AGB algorithm under channel estimation through closed-loop training [30] and investigation of nonlinear transmitter techniques with user scheduling =-=[31]-=- would be interesting direction to be investigated. Finally, we note that the proposed method can be nicely integrated into the dual codebooks structure in LTE-Advanced [32], [33] by feeding back the ... |