SCIENCE CHINA Information Sciences, Volume 62, Issue 2: 022303(2019) https://doi.org/10.1007/s11432-018-9413-6

## Impacts of practical channel impairments on the downlink spectral efficiency of large-scale distributed antenna systems

• AcceptedMar 29, 2018
• PublishedDec 18, 2018
Share
Rating

### Abstract

Channel impairments are major limiting factors in the performance of large-scale antenna systems.In this paper, we analyze the impacts of practical channel impairments caused by pilot contamination, Doppler shift, and phase noise on the downlink spectral efficiency of large-scale distributed antenna systems (L-DASs) with maximum ratio transmission (MRT) and zero-forcing (ZF) beamforming, in which per user power normalization is considered.Using a joint channel model that allows study of the simultaneous impacts of these channel impairments, we derive accurate and tractable closed-form approximations for the ergodic achievable downlink rate, thereby enabling spectral efficiency analysis of L-DASs and an efficient evaluation of the impacts of the channel impairments. It is shown that channel impairments reduce the downlink spectral efficiency and have a significant impact on ZF beamforming. The asymptotic user rate limit is also determined, from which we analyze the asymptotic performance of L-DASs with channel impairments. The analytical results show that MRT and ZF beamforming achieve the same asymptotic performance limit even with channel impairments. It is also found that the use of a large-scale antenna array at the base station sides can weaken the impacts of channel impairments.

### Acknowledgment

This work was supported in part by National Natural Science Foundation of China (NSFC) (Grant Nos. 61501113, 61571120, 61271205, 61521061, 61372100,), and Jiangsu Provincial Natural Science Foundation (Grant Nos. BK20150630, BK20151415).

### Supplement

Appendix

Proof of Theorem 1

For the signal power term $|{\rm~E}[{\boldsymbol~g}_{l,l,k}^{\text{H}}[t]~{\boldsymbol~w}_{l,k}^{\text{MRT}}[t]~]~|^2$, we have \begin{align} \big|{\rm E}\big[{\boldsymbol g}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{MRT}}[t] \big] \big|^2 \overset{({\rm a})}{=}&\, \big|{\rm E}\big[\|\hat{{\boldsymbol g}}_{l,l,k}[t] \|\big] \big|^2 \\ \overset{({\rm b})}=&\, \xi(\hat{k}_{l,l,k,\text{a}}[t])\hat{\theta}_{l,l,k,\text{a}}[t], \tag{39} \end{align} where $({\rm~a})$ is obtained because $\hat{{\boldsymbol~g}}_{l,l,k}[t]$ and $\tilde{{\boldsymbol~g}}_{l,l,k}[t]$ are independent, $({\rm~b})$ results from $\|\hat{{\boldsymbol~g}}_{l,l,k}[t]~\|~\sim~\text{Nakagami}(\hat{k}_{l,l,k,\text{a}}[t],\hat{k}_{l,l,k,\text{a}}[t]~\hat{\theta}_{l,l,k,\text{a}}[t])$ since $\|\hat{{\boldsymbol~g}}_{l,l,k}[t]~\|^2=|\hat{{\boldsymbol~g}}_{l,l,k}^{\text{H}}[t]{\boldsymbol~w}_{l,k}^{\text{MRT}}[t]|^2\sim~\Gamma(\hat{k}_{l,l,k,\text{a}}[t],~\hat{\theta}_{l,l,k,\text{a}}[t])$, which is obtained from Lemma 2 and (26).

Based on the independence of $\hat{{\boldsymbol~g}}_{l,l,k}[t]$ and considering the effect of pilot contamination, we decompose the interference power terms $\text{var}[{\boldsymbol~g}_{l,l,k}^{\text{H}}[t]{\boldsymbol~w}_{l,k}^{\text{MRT}}[t]]$, and $\sum_{(i,j)\neq~{(l,k)}}~{\rm~E}[~|{\boldsymbol~g}_{i,l,k}^{\text{H}}[t]~{\boldsymbol~w}_{i,j}^{\text{MRT}}[t]~|^2~]$ as follows: \begin{align} \text{var}\big[{\boldsymbol g}_{l,l,k}^{\text{H}}[t]{\boldsymbol w}_{l,k}^{\text{MRT}}[t]\big] = {\rm E}\big[ \big|\hat{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{MRT}}[t]\big|^2 \big] + {\rm E}\big[ \big|\tilde{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{MRT}}[t]\big|^2 \big] - \big|{\rm E}\big[\hat{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t]{\boldsymbol w}_{l,k}^{\text{MRT}}[t] \big] \big|^2, \tag{40} \end{align} and \begin{align} \sum\limits_{(i,j)\neq {(l,k)}} {\rm E}\big[ \big|{\boldsymbol g}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,j}^{\text{MRT}}[t] \big|^2 \big] =& \sum\limits_{j\neq k} {\rm E} \big[\big|{\boldsymbol g}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,j}^{\text{MRT}}[t] \big|^2 \big] + \sum\limits_{i\neq l}\sum\limits_{j\neq k} {\rm E} \big[\big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,j}^{\text{MRT}}[t] \big|^2 \big] \\ &+\sum\limits_{i\neq l}{\rm E} \big[\big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,k}^{\text{MRT}}[t] \big|^2\big] + \sum\limits_{i\neq l}\sum\limits_{j=1}^{K}{\rm E}\big[\big|\tilde{{\boldsymbol g}}_{i,l,k,}^{\text{H}}[t]{\boldsymbol w}_{i,j}^{\text{MRT}}[t] \big|^2\big]. \tag{41} \end{align} By applying Lemma 2 to approximate the distributions of the terms in $\left(\text{\ref{var}}~\right)$ and $\left(\text{\ref{com}}~\right)$, we have \begin{align} &\big|\hat{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{MRT}}[t]\big|^2 \sim \Gamma(\hat{k}_{l,l,k,\text{a}}[t],\hat{\theta}_{l,l,k,\text{a}}[t]), \tag{42} \\ &\big|\tilde{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{MRT}}[t]\big|^2 \sim \Gamma\left(\frac{1}{MN}\tilde{k}_{l,l,k,\text{a}}[t],\tilde{\theta}_{l,l,k,\text{a}}[t]\right), \tag{43} \\ &\big|{\boldsymbol g}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,j}^{\text{MRT}}[t] \big|^2 \sim \Gamma\left(\frac{1}{MN}k_{l,l,k,\text{a}}[t], \theta_{l,l,k,\text{a}}[t]\right), \tag{44} \\ &\big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,k}^{\text{MRT}}[t] \big|^2 \sim \Gamma(\hat{k}_{i,l,k,\text{a}}[t], \hat{\theta}_{i,l,k,\text{a}}[t]), \tag{45} \\ &\big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,j}^{\text{MRT}} [t]\big|^2 \sim \Gamma\left(\frac{1}{MN}\hat{k}_{i,l,k,\text{a}}[t], \hat{\theta}_{i,l,k,\text{a}}[t]\right), \tag{46} \\ &\big|\tilde{{\boldsymbol g}}_{i,l,k,}^{\text{H}}[t]{\boldsymbol w}_{i,j}^{\text{MRT}}[t] \big|^2 \sim \Gamma\left(\frac{1}{MN}\tilde{k}_{i,l,k,\text{a}}[t], \tilde{\theta}_{i,l,k,\text{a}}[t]\right). \tag{47} \end{align} Substituting (39) and (42)–(47) into (19) concludes the proof.

Proof of Theorem 2

First, given the distributions of $\hat{{\boldsymbol~g}}_{i,l,k}^{\text{H}}[t]~\hat{{\boldsymbol~g}}_{i,l,k}[t]$ in (26), the non-isotropic achievable CSI $\hat{{\boldsymbol~g}}_{i,l,k}[t]$ can be approximated as an isotropic vector $\hat{{\boldsymbol~g}}_{i,l,k,\text{a}}[t]$ with i.i.d. $\mathcal{CN}(0,\hat{\theta}_{i,l,k,\text{a}})$ elements [29]. Then, with the definition of ${\boldsymbol~F}_{l,\text{a}}[t]\triangleq[\hat{{\boldsymbol~g}}_{l,l,1,\text{a}}[t],~\ldots,$$\hat{{\boldsymbol~g}}_{l,l,K,\text{a}}[t]~]$, the useful signal power term $|{\rm~E}[{\boldsymbol~g}_{l,l,k}^{\text{H}}[t]~{\boldsymbol~w}_{l,k}^{\text{ZF}}[t]~]~|^2$ can be calculated by \begin{align} \big|{\rm E}\big[{\boldsymbol g}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{ZF}}[t] \big] \big|^2 &\overset{({\rm a})}{=} \big|{\rm E}[ 1 / \|{\boldsymbol f}_{l,l,k}[t] \|] \big|^2 \\ & \overset{({\rm b})}= \Big|{\rm E}\Big[\big( \big[\big({\boldsymbol F}_{l}^{\text{H}}[t] {\boldsymbol F}_l[t] \big)^{-1} \big]_{k,k} \big)^{-1/2} \Big] \Big|^2 \\ & \overset{({\rm c})}= \xi(\rho\hat{k}_{l,l,k,\text{a}}[t])\hat{\theta}_{l,l,k,\text{a}}[t], \tag{48} \end{align} where $({\rm~a})$ is obtained because of the independence of ${\boldsymbol~w}_{l,k}^{\text{ZF}}[t]$ and $\tilde{{\boldsymbol~g}}_{l,l,k}[t]$ and $\hat{{\boldsymbol~g}}_{l,l,k}^{\text{H}}[t]{\boldsymbol~w}_{l,k}^{\text{ZF}}[t]=1/\|{\boldsymbol~f}_{l,l,k}[t]~\|$, $({\rm~b})$ results from $\|{\boldsymbol~f}_{l,l,k}[t]\|^2~=~[({\boldsymbol~F}_{l}^{\text{H}}[t]~{\boldsymbol~F}_l[t]~)^{-1}~]_{k,k}~$, $({\rm~c})$ results from $1/\|{\boldsymbol~f}_{l,l,k}[t]\|\sim~\text{Nakagami}(\rho\hat{k}_{l,l,k,\text{a}}[t],~\rho\hat{k}_{l,l,k,\text{a}}[t]~\hat{\theta}_{l,l,k,\text{a}}[t])$ where we have applied Lemma 2 to approximate the distribution of $[({\boldsymbol~F}_{l}^{\text{H}}[t]~{\boldsymbol~F}_l[t]~)^{-1}~]_{k,k}$ as $\Gamma(\rho\hat{k}_{l,l,k,\text{a}}[t],~\hat{\theta}_{l,l,k,\text{a}}[t]~)$ since $[({\boldsymbol~F}_{l,\text{a}}^{\text{H}}[t]~{\boldsymbol~F}_{l,\text{a}}[t]~)^{-1}~]_{k,k}$ $\sim~\Gamma(MN-K+1,~\hat{\theta}_{l,l,k,\text{a}}[t]~)$ 1).Similar to the analysis in the proof given for Theorem 1 and considering that $\hat{{\boldsymbol~g}}_{l,l,k}[t]~{\boldsymbol~w}_{l,j}^{\text{ZF}}[t]~=~0$ for $j\neq~k$, the interference power terms $\text{var}[{\boldsymbol~g}_{l,l,k}^{\text{H}}[t]{\boldsymbol~w}_{l,k}^{\text{ZF}}[t]]$ and $\sum_{(i,j)\neq~{(l,k)}}~{\rm~E}[~|{\boldsymbol~g}_{i,l,k}^{\text{H}}[t]~{\boldsymbol~w}_{i,j}^{\text{ZF}}[t]~|^2~]$ can be decomposed as \begin{align} \text{var}\big[{\boldsymbol g}_{l,l,k}^{\text{H}}[t]{\boldsymbol w}_{l,k}^{\text{ZF}}[t]\big] ={\rm E}\big[ \big|\hat{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{ZF}}[t]\big|^2 \big] + {\rm E}\big[ \big|\tilde{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{ZF}}[t]\big|^2 \big] - \big|{\rm E}\big[\hat{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t]{\boldsymbol w}_{l,k}^{\text{ZF}}[t] \big] \big|^2, \tag{49} \end{align} and \begin{align} \sum_{(i,j)\neq {(l,k)}} {\rm E}\big[ \big|{\boldsymbol g}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,j}^{\text{ZF}}[t] \big|^2 \big] =& \sum_{j\neq k} {\rm E} \big[\big|\tilde{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,j}^{\text{ZF}}[t] \big|^2 \big]+ \sum_{i\neq l}\sum_{j\neq k} {\rm E} \big[\big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,j}^{\text{ZF}}[t] \big|^2 \big] \\ &+ \sum_{i\neq l}{\rm E} \big[\big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,k}^{\text{ZF}}[t] \big|^2 \big]+ \sum_{i\neq l}\sum_{j=1}^{K}{\rm E}\big[\big|\tilde{{\boldsymbol g}}_{i,l,k,}[t]{\boldsymbol w}_{i,j}^{\text{ZF}} \big|^2\big]. \tag{50} \end{align} The distributions of the terms in $\left(\text{\ref{var_zf}}~\right)$ and $\left(\text{\ref{inf2c_zf}}~\right)$ can be obtained by applying Lemma 2, \begin{align} \big|\hat{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{ZF}}[t]\big|^2& \sim\Gamma(\rho\hat{k}_{l,l,k,\text{a}}[t],\hat{\theta}_{l,l,k,\text{a}}[t]), \tag{51} \\ \big|\tilde{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,k}^{\text{ZF}}[t]\big|^2 &\sim \Gamma\left(\frac{1}{MN}\tilde{k}_{l,l,k,\text{a}}[t],\tilde{\theta}_{l,l,k,\text{a}}[t]\right), \tag{52} \\ \big|\tilde{{\boldsymbol g}}_{l,l,k}^{\text{H}}[t] {\boldsymbol w}_{l,j}^{\text{ZF}}[t] \big|^2& \sim\Gamma\left(\frac{1}{MN}\tilde{k}_{l,l,k,\text{a}}[t],\tilde{\theta}_{l,l,k,\text{a}}[t]\right), \tag{53} \\ \big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,k}^{\text{ZF}}[t] \big|^2&\sim \Gamma \big(\rho\hat{k}_{i,l,k,\text{a}}[t], \hat{\theta}_{i,l,k,\text{a}}[t]\big), \tag{54} \\ \big|\hat{{\boldsymbol g}}_{i,l,k}^{\text{H}}[t] {\boldsymbol w}_{i,j}^{\text{ZF}}[t] \big|^2&\sim \Gamma\left(\frac{1}{MN}\hat{k}_{i,l,k,\text{a}}[t], \hat{\theta}_{i,l,k,\text{a}}[t]\right), \tag{55} \\ \big|\tilde{{\boldsymbol g}}_{i,l,k}[t]{\boldsymbol w}_{i,j}^{\text{ZF}}[t] \big|^2&\sim \Gamma\left(\frac{1}{MN}\tilde{k}_{i,l,k,\text{a}}[t],\tilde{\theta}_{i,l,k,\text{a}}[t]\right). \tag{56} \end{align} Substituting (48) and (51)–(56) into (19) yields the closed-form approximation (34). This completes the proof.

Khansefid A, Minn H. Achievable downlink rates of MRC and ZF precoders in massive MIMO with uplink and downlink pilot contamination. IEEE Trans Commun, 2015, 63: 4849–4864.

### References

[1] Marzetta T L. Noncooperative Cellular Wireless with Unlimited Numbers of Base Station Antennas. IEEE Trans Wireless Commun, 2010, 9: 3590-3600 CrossRef Google Scholar

[2] Wang D, Zhang Y, Wei H. An overview of transmission theory and techniques of large-scale antenna systems for 5G wireless communications. Sci China Inf Sci, 2016, 59: 081301 CrossRef Google Scholar

[3] Lu L, Li G Y, Swindlehurst A L. An Overview of Massive MIMO: Benefits and Challenges. IEEE J Sel Top Signal Process, 2014, 8: 742-758 CrossRef ADS Google Scholar

[4] Rusek F, Persson D, Lau B K, et al. Scaling up MIMO: opportunities and challenges with very large arrays. IEEE Signal Process Mag, 2013, 30: 40--60. Google Scholar

[5] Zhang J, Wen C K, Jin S. On Capacity of Large-Scale MIMO Multiple Access Channels with Distributed Sets of Correlated Antennas. IEEE J Sel Areas Commun, 2013, 31: 133-148 CrossRef Google Scholar

[6] Fernandes F, Ashikhmin A, Marzetta T L. Inter-Cell Interference in Noncooperative TDD Large Scale Antenna Systems. IEEE J Sel Areas Commun, 2013, 31: 192-201 CrossRef Google Scholar

[7] Adhikary A, Ashikhmin A, Marzetta T L. Uplink interference reduction in large scale antenna systems. In: Proceedings of IEEE International Symposium on Information Theory, Honolulu, 2014. 2529--2533. Google Scholar

[8] Wen C K, Jin S, Wong K K. Channel Estimation for Massive MIMO Using Gaussian-Mixture Bayesian Learning. IEEE Trans Wireless Commun, 2015, 14: 1356-1368 CrossRef Google Scholar

[9] Truong K T, Heath R W. Effects of channel aging in massive MIMO systems. J Commun Netw, 2013, 15: 338-351 CrossRef Google Scholar

[10] Papazafeiropoulos A K, Ngo H Q, Matthaiou M, et al. Uplink performance of conventional and massive MIMO cellular systems with delayed CSIT. In: Proceedings of IEEE International Symposium on Personal, Indoor, and Mobile Radio Communication (PIMRC), Washington, 2014. 601--606. Google Scholar

[11] Papazafeiropoulos A K, Ratnarajah T. Deterministic Equivalent Performance Analysis of Time-Varying Massive MIMO Systems. IEEE Trans Wireless Commun, 2015, 14: 5795-5809 CrossRef Google Scholar

[12] Papazafeiropoulos A K. Impact of user mobility on optimal linear receivers in cellular networks. In: Proceedings of IEEE International Conference on Communications (ICC), London, 2015. 2239--2244. Google Scholar

[13] Guo K, Khodapanah B, Ascheid G. Performance analysis of downlink MMSE beamforming training in TDD MU-massive-MIMO. In: Proceedings of IEEE Wireless Communications and Networking Conference (WCNC), Doha, 2016. Google Scholar

[14] Pitarokoilis A, Moammed S K, Larsson E G. Achievable rates of ZF receivers in massive MIMO with phase noise impairments. In: Proceedings of Asilomar Conference on Signals, Systems and Computers, Pacific Grove, 2013.łinebreak 1004--1008. Google Scholar

[15] Pitarokoilis A, Mohammed S K, Larsson E G. Uplink Performance of Time-Reversal MRC in Massive MIMO Systems Subject to Phase Noise. IEEE Trans Wireless Commun, 2015, 14: 711-723 CrossRef Google Scholar

[16] Bjornson E, Matthaiou M, Debbah M. Massive MIMO with Non-Ideal Arbitrary Arrays: Hardware Scaling Laws and Circuit-Aware Design. IEEE Trans Wireless Commun, 2015, 14: 4353-4368 CrossRef Google Scholar

[17] Corvaja R, Armada A G. Phase Noise Degradation in Massive MIMO Downlink With Zero-Forcing and Maximum Ratio Transmission Precoding. IEEE Trans Veh Technol, 2016, 65: 8052-8059 CrossRef Google Scholar

[18] Zhu J, Schober R, Bhargava V K. Physical layer security for massive MIMO systems impaired by phase noise. In: Proceedings of the 17th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Edinburgh, 2016. Google Scholar

[19] Larsson E, Edfors O, Tufvesson F, et al. Massive MIMO for next generation wireless systems. IEEE Commun Mag, 2014, 52: 184--195. Google Scholar

[20] Lee S R, Moon S H, Kim J S. Capacity analysis of distributed antenna systems in a composite fading channel. IEEE Trans Wireless Commun, 2012, 11: 1076-1086 CrossRef Google Scholar

[21] Zhu P, You X, Li J. Spectral efficiency analysis of large-scale distributed antenna system in a composite correlated Rayleigh fading channel. IET Commun, 2015, 9: 681-688 CrossRef Google Scholar

[22] Wang D M, You X H, Wang J Z, et al. Spectral efficiency of distributed MIMO cellular systems in a composite fading channel. In: Proceedings of IEEE International Conference on Communications (ICC'08), Prague, 2008. 1259--1264. Google Scholar

[23] Wang J, Dai L. Asymptotic Rate Analysis of Downlink Multi-User Systems With Co-Located and Distributed Antennas. IEEE Trans Wireless Commun, 2015, 14: 3046-3058 CrossRef Google Scholar

[24] Wang J, Dai L. Downlink Rate Analysis for Virtual-Cell Based Large-Scale Distributed Antenna Systems. IEEE Trans Wireless Commun, 2016, 15: 1998-2011 CrossRef Google Scholar

[25] Björnson E, Matthaiou M, Pitarokoilis A, et al. Distributed massive MIMO in cellular networks: impact of imperfect hardware and number of oscillators. In: Proceedings of European Signal Processing Conference (EUSIPCO), Nice, 2015. 2436--2440. Google Scholar

[26] Papazafeiropoulos A K. Impact of General Channel Aging Conditions on the Downlink Performance of Massive MIMO. IEEE Trans Veh Technol, 2017, 66: 1428-1442 CrossRef Google Scholar

[27] Li J M, Wang D M, Zhu P C, et al. Uplink spectral efficiency analysis of distributed massive MIMO with channel impairments. IEEE Access, 2017, 5: 5020--5030. Google Scholar

[28] Interdonato G, Ngo H Q, Larsson E G, et al. How much do downlink pilots improve cell-free massive MIMO? In: Proceedings of IEEE Global Communications Conference (GLOBECOM), Washington, 2016. Google Scholar

[29] Li J, Wang D, Zhu P. Downlink Spectral Efficiency of Distributed Massive MIMO Systems With Linear Beamforming Under Pilot Contamination. IEEE Trans Veh Technol, 2018, 67: 1130-1145 CrossRef Google Scholar

[30] Jakes W C. Microwave Mobile Communications. New York: Wiley, 1974. Google Scholar

[31] Krishnan R, Khanzadi M R, Krishnan N. Linear Massive MIMO Precoders in the Presence of Phase Noise-A Large-Scale Analysis. IEEE Trans Veh Technol, 2016, 65: 3057-3071 CrossRef Google Scholar

[32] Carvalho E de, Björnson E, Larsson E G, et al. Random access for massive MIMO systems with intra-cell pilot contamination. In: Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Shanghai, 2016. 3361--3365. Google Scholar

[33] Truong K T, Lozano A, Heath R W, et al. Optimal training in continuous flat-fading massive MIMO systems. In: Proceedings of European Wireless Conference, Barcelona, 2014. Google Scholar

[34] Kay S M. Fundamental of Statistical Signal Processing: Estimation Theory. Englewood: Prentice-Hall, 1993. Google Scholar

[35] Jose J, Ashikhmin A, Marzetta T L. Pilot Contamination and Precoding in Multi-Cell TDD Systems. IEEE Trans Wireless Commun, 2011, 10: 2640-2651 CrossRef Google Scholar

[36] Björnson E, Larsson E G, Marzetta T L. Massive MIMO: ten myths and one critical question. IEEE Commun Mag, 2016, 54: 114--123. Google Scholar

[37] Björnson E, Larsson E G, Debbah M. Massive MIMO for maximal spectral efficiency: how many users and pilots should be allocated? IEEE Trans Wirel Commun, 2016, 15: 1293--1308. Google Scholar

[38] Van Chien T, Bjornson E, Larsson E G. Joint Power Allocation and User Association Optimization for Massive MIMO Systems. IEEE Trans Wireless Commun, 2016, 15: 6384-6399 CrossRef Google Scholar

[39] Hoydis J, Brink S ten, Debbah M. Massive MIMO in the UL/DL of cellular networks: how many antennas do we need? IEEE J Sel Areas Commun, 2013, 31: 160--171. Google Scholar

[40] Kammoun A, Muller A, Bjornson E. Linear Precoding Based on Polynomial Expansion: Large-Scale Multi-Cell MIMO Systems. IEEE J Sel Top Signal Process, 2014, 8: 861-875 CrossRef ADS arXiv Google Scholar

[41] Li J M, Wang D M, Zhu P C, et al. Downlink spectral efficiency of multi-cell multi-user large-scale DAS with pilot contamination. In: Proceedings of IEEE International Conference on Communications (ICC), London, 2015.łinebreak 2011--2016. Google Scholar

[42] Heath Jr R W, Wu T, Kwon Y H. Multiuser MIMO in Distributed Antenna Systems With Out-of-Cell Interference. IEEE Trans Signal Process, 2011, 59: 4885-4899 CrossRef ADS Google Scholar

[43] Zhang J, Andrews J G. Adaptive Spatial Intercell Interference Cancellation in Multicell Wireless Networks. IEEE J Sel Areas Commun, 2010, 28: 1455-1468 CrossRef Google Scholar

[44] Hosseini K, Yu W, Adve R S. Large-Scale MIMO Versus Network MIMO for Multicell Interference Mitigation. IEEE J Sel Top Signal Process, 2014, 8: 930-941 CrossRef ADS arXiv Google Scholar

[45] Hosseini K, Yu W, Adve R S. Modeling and analysis of ergodic capacity in network MIMO systems. In: Proceedings of IEEE Globecom Workshops (GC Wkshps), Austin, 2014. 808--814. Google Scholar

[46] Hosseini K, Yu W, Adve R S. A Stochastic Analysis of Network MIMO Systems. IEEE Trans Signal Process, 2016, 64: 4113-4126 CrossRef ADS arXiv Google Scholar

[47] Seifi N, Heath R W, Coldrey M, et al. Joint transmission mode and tilt adaptation in coordinated small-cell networks. In: Proceedings of IEEE International Conference on Communications Workshops (ICC), Sydney, 2016. 598--603. Google Scholar

• Figure 1

(Color online) Fluctuation of the norms of MRT beamforming vectors around their means versus the total number of transmit antennas.

• Figure 2

(Color online) Cumulative distribution function of the average achievable rate per user at time $t=1$ for $L=7$, $T_{\text{s}}=10^{-5}$ s, $T_{\text{c}}=100$, $\sigma_{\varphi_{l,k}}=\sigma_{\phi_{i,m}}=0.72^{\circ}$ with different $K$, $M$ and $N$.

• Figure 3

(Color online) Average achievable rate per user versus phase noise increment standard deviations for different values of $f_{\text{D}}T_{\text{s}}$, $L=7$, $M=10$, $N=20$, $K=8$, $T_{\text{s}}=10^{-5}$ s, $T_{\text{c}}=100$.

• Figure 4

(Color online) Average achievable rate per user with channel impairments versus the total number of transmit antennas, $T_{\text{c}}=100$, $L=7$, $M=5$, $K=8$.

Citations

• #### 0

Altmetric

Copyright 2020 Science China Press Co., Ltd. 《中国科学》杂志社有限责任公司 版权所有