Beam Domain Channel Estimation for Spatial Non-Stationary Massive MIMO Systems
Lin Hou, Hengtai Chang, Cheng-Xiang Wang, Jie Huang, Songjiang Yang · 2025
Read Beam Domain Channel Estimation for Spatial Non-Stationary Massive MIMO Systems by Lin Hou, Hengtai Chang, Cheng-Xiang Wang, Jie Huang, Songjiang Yang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-04242.txt.
Introduction
Lin Hou and four co-authors tackle channel estimation in massive antenna systems in a 2025 paper. In such systems the radio channel changes across space, and energy leaks between beams, which breaks simple estimators. They first build a realistic beam-domain model that captures both effects through the idea of a visibility region, the part of the array each path can actually reach.
On that model they design an estimation method that exploits the channel's block-like sparse pattern across beams. A power threshold guides which blocks to trust. Simulations show the method estimates accurately with few pilot signals and modest computing cost, beating conventional schemes in their comparisons.
Catalog record
| CREATOR | Lin Hou, Hengtai Chang, Cheng-Xiang Wang, Jie Huang, Songjiang Yang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-04242
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Beam Domain Channel Estimation for Spatial Non-Stationary Massive MIMO Systems Lin Hou, Hengtai Chang, Cheng-Xiang Wang, Jie Huang, Songjiang Yang Abstract In massive multiple-input multiple-output (MIMO) systems, the channel estimation scheme is subject to the spatial non-stationarity and inevitably power leakage in the beam domain. In this paper, a beam domain channel estimation scheme is investigated for spatial non-stationary (SNS) massive MIMO systems considering power leakage. Specifically, a realistic massive MIMO beam domain channel model (BDCM) is introduced to capture the spatial non-stationarity considering power leakage by introducing the illustration of visibility region (VR). Then, a beam domain structure-based sparsity adaptive matching pursuit (BDS-SAMP) scheme is proposed based on the cross-block sparse structure and power ratio threshold of beam domain channel. Finally, the simulation results validate the accuracy of proposed BDS-SAMP scheme with low pilot overhead and reasonable complexity by comparing with conventional schemes. I Introduction In recent years, massive MIMO has attracted wide attention due to the improvement in efficiency of spectrum and energy [ 1 , 2 ] . However, the computational complexity of channel modeling and the overhead of channel estimation are challenging task because of the large number of antennas [ 4 , 3 ] . The BDCM, transferring the channel model from array domain into the beam domain, can reduce the computational complexity and play a critical role in the design and performance evaluation [ 5 ] . In order to fully maintain the beamforming gains of massive MIMO beam domain channels, it is essential to obtain accurate knowledge of channel state information (CSI). Massive MIMO beam domain channels display sparsity with a small number of channel elements containing most of the energy [ 6 ] . Numerous channel estimation algorithms based on the compressed sensing (CS) [ 7 ] including convex optimization algorithms [ 8 ] , Bayesian algorithms [ 9 ] , and greedy algorithms [ 10 ] have been employed for massive MIMO systems by leveraging the inherent sparsity of channels. Due to the advantages in terms of implementation simplicity and computational efficiency, greedy algorithms such as orthogonal matching pursuit (OMP), sparse adaptive matching pursuit (SAMP) [ 11 ] , and block-based OMP (BOMP) [ 12 ] have been used for massive MIMO channel estimation. Nevertheless, the above-mentioned algorithms considered massive MIMO channels to be spatial stationary (SS) with identical sparse structure. In practice, massive MIMO channels are SNS with some scatterers visible only to a portion of the antenna array [ 13 ] , resulting in varied sparse structure over array. In [ 14 ] and [ 15 ] , the authors modified the standardized 3GPP channel model to capture the spatial non-stationarity by using birth-death process, and then performed array domain channel estimation for SNS massive MIMO systems. However, they ignored the inevitably impact of channel power leakage on the sparsity and estimation performance since massive MIMO channels are not ideally sparse in practice. In [ 16 ] and [ 17 ] , the channel estimation schemes based on support detection (SD) [ 18 ] , e.g., adaptive SD (ASD) and successive SD, were proposed by making full use of adaptive updated or successive rectangle support structure of beam domain channel considering power leakage. However, these schemes were based on SS massive MIMO channels assumption, which will further underestimate the power leakage of beam domain channel and cause inaccurate results of channel estimation. To the best of our knowledge, beam domain channel estimation for SNS massive MIMO systems is still lacking in the literature. To fill this research gap, a BDS-SAMP channel estimation scheme is proposed for SNS massive MIMO systems considering sparse structure and power leakage of beam domain channel. The main contributions and novelties of this paper are summarized as below. • A realistic massive MIMO BDCM is introduced to capture spatial non-stationarity by introducing the VR of different clusters. Furthermore, the beam sparse structure and power leakage effect are demonstrated. • A BDS-SAMP channel estimation scheme for SNS massive MIMO systems is proposed, in which the beam support set is iteratively obtained based on the cross-block sparse structure and power ratio threshold of beam domain channel. • Simulation results verify the accuracy and robustness of proposed BDS-SAMP scheme by comparing with the traditional channel estimation schemes. The remainder of this paper is organized as follows. Section II introduces the system model including SNS massive MIMO BDCM and problem formulation. Based on it, Section III analyzes the power leakage effect. In section IV, a BDS-SAMP channel estimation scheme for SNS massive MIMO systems is proposed. Section V presents the simulation results and demonstrates the predominant performance of the proposed scheme. Finally, conclusions are drawn in Section VI. II System Model Channel estimation schemes are highly dependent on channel models. The realistic channel model is the basis for system performance evaluation and accurate channel estimation. In this section, a realistic SNS massive MIMO BDCM is introduced by introducing the illustration of VR. Based on it, the beam domain channel estimation problem is formulated. II-A SNS Massive MIMO BDCM We consider a downlink massive MIMO system as shown in Fig. 1 , where the base station (BS) employs a large uniform planar array (UPA) with {P} antennas and serves U single-antenna users. Here, P consists of {P_{v}} antenna elements in vertical dimension multiplied by {P_{h}} antenna elements in horizontal dimension. Fig. 1: System structure diagram for massive MIMO systems. We classify all clusters in massive MIMO systems into wholly visible (WV) clusters and partially visible (PV) clusters [ 5 ] . Each cluster has a corresponding VR, the VR of WV clusters is the entire array, while that of PV clusters is the partial array. The ratio of PV clusters to total clusters is \rho . We start with the geometry-based stochastic channel model (GBSM) for massive MIMO to obtain array domain channel transfer function (CTF). The CTF matrix for the u th (u=1,2,...,U) user can be written as \displaystyle{\mathbf{H}_{u}}={{\mathbf{H}}_{u}^{{\text{WV}}}}+{{\mathbf{H}}_{u}^{{\text{PV}}}} (1) where {\mathbf{H}}_{u}^{{\text{WV}}} and {\mathbf{H}}_{u}^{{\text{PV}}} are CTF matrices of WV and PV multipath components (MPCs), which can be formulated as \displaystyle{\mathbf{H}}_{u}^{{\text{WV}}}=\sum\limits_{n\in{{\mathbf{N}}_{1}}}{\sum\limits_{m=1}^{{M_{n}}}{{\beta_{n,m}}{e^{{\rm j}\psi_{n,m}}}}}{{{\mathbf{U}}\left({\theta_{n,m}^{az},\theta_{n,m}^{el}}\right)}} (2) \displaystyle{{\mathbf{H}}_{u}^{{\text{PV}}}}=\sum\limits_{n\in{{\mathbf{N}}_{2}}}{\sum\limits_{m=1}^{{M_{n}}}{{\beta_{n,m}}{e^{{\rm j}\psi_{n,m}}}}}{{\mathbf{\hat{U}}}\left({\theta_{n,m}^{az},\theta_{n,m}^{el}}\right)} (3) with \psi_{n,m}={-2\pi f{\tau_{n,m}}+{\varphi_{n,m}}}. (4) Here, f indicates the carrier frequency, M_{n} stands for the paths number of the n th cluster, {\bf{N}}_{1/2} are the sets of the WV/PV clusters, \varphi_{n,m} , \tau_{n,m} and {\beta_{n,m}} denote the random initial phase, delay and coefficient of m th path in the n th cluster, respectively. To avoid redundancy, abbreviate all variables whose subscript is ( {n,m} ) as that of m_{n} th path in the following part. The UPA steering matrix for WV clusters is defined as {\bf{U}}\left({\theta_{n,m}^{az},\theta_{n,m}^{el}}\right)={\left[{{\mathbf{b}}\left({\theta_{n,m}^{el}}\right)\otimes{\mathbf{a}}\left({\theta_{n,m}^{az}}\right)}\right]}^{\text{T}} , where {\mathbf{a}}\left({\theta_{n,m}^{az}}\right)={\left[{1,{e^{{\rm j}2\pi\theta_{n,m}^{az}}},\cdots,{e^{{\rm j}2\pi\left({{P_{h}}-1}\right)\theta_{n,m}^{az}}}}\right]^{\text{T}}} and {\mathbf{b}}\left({\theta_{n,m}^{el}}\right)={\left[{1,{e^{{\rm j}2\pi\theta_{n,m}^{el}}},\cdots,{e^{{\rm j}2\pi\left({{P_{v}}-1}\right)\theta_{n,m}^{el}}}}\right]^{\text{T}}} with \theta_{n,m}^{az} and \theta_{n,m}^{el} standing for the spatial frequencies in azimuth and elevation directions associated with the m_{n} th path. Without loss of generality, we usually set adjacent antennas spacing as half wavelength, thus \theta_{n,m}^{az/el}\in\left[{-0.5,0.5}\right] . Furthermore, the steering matrix for PV clusters can be obtained via {\bf{\hat{U}}}\left({\theta_{n,m}^{az},\theta_{n,m}^{el}}\right)={\bf{U}}\left({\theta_{n,m}^{az},\theta_{n,m}^{el}}\right)\odot{{\mathbf{\xi}}_{n,m}} . The VR of m_{n} th path is defined by the matrix \mathbf{\xi}_{n,m} consisting only of 0 and 1, which indicates the visibility of cluster to antenna array. Fig. 2: Flowchart of generating the SNS BDCM. The BDCMs can be obtained by