Pilot length optimization for spatially correlated multi-user MIMO channel estimation

Abstract

We address the design of pilot sequences for channel estimation in the context of multiple-user Massive MIMO; considering the presence of channel correlation, and assuming that the statistics are known, we seek to exploit the spatial correlation of the channels to minimize the length of the pilot sequences, and specifically the fact that the users can be separated either through their spatial signature (low-rank channel covariance matrices), or through the use of different training sequences. We introduce an algorithm to design short training sequences for a given set of user covariance matrices. The obtained pilot sequences are in general non-orthogonal, however they ensure that the channel estimation error variance is uniformly upper-bounded by a chosen constant over all channel dimensions. We show through simulations using a realistic scenario based on the one-ring channel model that the proposed technique can yield pilot sequences of length significantly smaller than the number of users in the system.

Performance Scaling Law for Multicell Multiuser Massive MIMO

“…For more practical channel models, by exploiting the large spatial degrees of freedom provided by the massive antenna array at the BS, several efficient methods have been proposed to reduce pilot contamination [20]- [23]. Specifically, the difference among users' channel directions [20] or locations [21] and smart pilot allocation based on them [22] were utilized to avoid pilot contamination, and corresponding simulations showed that as M increases, lower pilot contamination level can be achieved. In our work, the effect of PCE is reflected by the L p parameter in our modeling, which is assumed to be a constant.…”

Covariance estimation with projected data: Applications to CSI covariance acquisition and tracking

“…The assumed channel models are generally such that statistical CSI can be identified to the second-order statistics, through a spatial covariance matrix associated to each channel state process. In particular, an abundant literature is dedicated to the topic of reducing the amount of reference symbols dedicated to the estimation of instantaneous CSI in multi-user systems, for which a variety of approaches have been proposed [4]- [11]; all these techniques have in common the idea that the prior information contained in the statistics can substantially reduce the amount of reference symbols dedicated to CSI estimation, by allowing either a denser reuse of identical pilot sequences, or the use of non-orthogonal pilot sequences without sacrificing estimation accuracy.…”

“…Although a vast statistical literature on covariance estimation is available, focusing among others on large dimensional analysis [12], many of the existing approaches do not directly apply to the problem at hand if one takes into account the dynamic user scheduling and pilot sequence allocation (see [4]- [11]) resulting from evolving channel statistics.…”

SNOPS: Short Non-Orthogonal Pilot Sequences for Downlink Channel State Estimation in FDD Massive MIMO

“…In [13], the design of non-orthogonal pilots based on per-user spatial covariance information has been addressed for uplink multi-user Massive MIMO CSI estimation; the object of the present article is to introduce a downlink counterpart to this approach. In the sequel, we seek to optimize the design of the training sequences transmitted by the BTS during downlink CSI estimation.…”