Estimation of Wideband Dynamic mmWave and THz Channels for 5G Systems and Beyond

Abstract

Millimeter wave (mmWave) wideband channels in a multiple-input multiple-output (MIMO) transmission are described by a sparse set of impulse responses in the angle-delay, or space-time (ST), domain. These characteristics will be even more prominent in the THz band used in future systems. We consider two approaches for channel estimation: compressedsensing (CS), exploiting the sparsity in the angular/delay domain, and low-rank (LR), exploiting the algebraic structure of channel matrix. Both approaches share several commonalities, and this paper provides for the first time i) a comparison of the two approaches, and ii) new versions of CS and LR methods that significantly improve performance in terms of mean squared error (MSE), computational complexity, and latency. We derive the asymptotic MSE bound for any estimator of the ST-MIMO multipath channels with invariant angles/delays and time-varying fading, with unknown angle/delay diversity order: the bound also accounts for the degradation introduced by sub-optimal separable channel models. We will show that in the considered scenarios both CS and LR approaches attain the bound. Our performance assessment over ideal and 3 rd generation partnership project (3GPP) channel models, suitable for the fifth-generation (5G) and beyond of cellular networks, shows the trade-off obtained by the methods over various metrics: i) CS methods are converging faster than the LR methods, both attaining the asymptotic MSE bound; ii) the CS methods depend on the array manifold, while LR methods are independent of the array calibration; iii) CS solutions are more complex than LR solutions.

Channel Estimation for 6G V2X Hybrid Systems Using Multi-Vehicular Learning

“…Extensive numerical simulations, based on ray-tracinggenerated channel data [35] and realistic vehicle trajectories [36], show that the proposed LR channel estimation method outperforms the Unconstrained Maximum Likelihood (U-ML) in terms of Mean Square Error (MSE) on channel estimation and Spectral Efficiency (SE) for both FC-HBF and SC-HBF architectures. Furthermore, we prove numerically that the MSE of the LR attains the theoretical MSE lower bound derived in [28] for structured MIMO channels as considered here. In particular, for a target SE, both architectures and LR methods achieve an SNR gain up to 15 dB in single-path and up to 10 dB in multipath scenarios.…”

“…In particular, the work in [28] demonstrates that LR methods attain similar performances to CS with lower sensitivity to hardware impairments. In [28], the LR channel estimation is enabled by consecutive transmissions of training blocks, that limits the application to static or low-mobility scenarios.…”

“…Moreover, the joint optimization of both analog and digital precoders/combiners increases the complexity and the cost of the implementation in practical high-mobility systems, as the channel is rapidly time-varying. Finally, as any grid-based technique, CS has a significant drawback in the high sensitivity to array calibrations [28], which is critical in hybrid systems [29].…”

“…Originally proposed in [30]- [33] for lower-spectrum systems, where the channel is not sufficiently sparse to boost the LR application to practical systems, the LR has recently been resumed for mmW/sub-THz systems, for Fully Digital (FD) systems only [28], [34]. In particular, the work in [28] demonstrates that LR methods attain similar performances to CS with lower sensitivity to hardware impairments. In [28], the LR channel estimation is enabled by consecutive transmissions of training blocks, that limits the application to static or low-mobility scenarios.…”

“…can be retrieved from the single received training signal Y as the combination of a training sequence-dependent matrix G and another one referred as position-dependent matrix Π(p,θ), that follows a derivation similar to [28] (not Analog Combiner DFT Indexes…”