Abstract
This paper proposes a novel direction-of-arrival (DOA)-aided channel estimation for a hybrid millimeter-wave (mm-wave) massive multiple-input multiple-output system with a uniform planar array at the base station. To explore the physical characteristics of the antenna array in mm-wave systems, the parameters of each channel path are decomposed into the DOA information and the channel gain information. We first estimate the initial DOAs of each uplink path through the 2-D discrete Fourier transform and enhance the estimation accuracy via the angle rotation technique. We then estimate the channel gain information using a small amount of training resources, which significantly reduces the training overhead and the feedback cost. More importantly, to examine the estimation performance, we derive the theoretical bounds of the mean squared errors (MSEs) and the Cramer–Rao lower bounds (CRLBs) of the joint DOA and channel gain estimation. The simulation results show that the performances of the proposed methods are close to the theoretical MSEs’ analysis. Furthermore, the theoretical MSEs are also close to the corresponding CRLBs.
Highlights
A S AN important candidate in the fifth generation (5G) mobile communications, the millimeter-wave communication that explores large amount of bandwidth resources at frequencies 30GHz to 300GHz has been proposed for outdoor cellular systems [1]–[4]
We focus on the DOA estimation and channel estimation for the mm-wave massive multipleinput multiple-output (MIMO) system with hybrid precoding
We proposed a novel channel estimation for hybrid digital and analog mm-wave massive MIMO system, where the channel is decomposed into DOA information and channel gain information
Summary
A S AN important candidate in the fifth generation (5G) mobile communications, the millimeter-wave (mm-wave) communication that explores large amount of bandwidth resources at frequencies 30GHz to 300GHz has been proposed for outdoor cellular systems [1]–[4]. Many high resolution subspace based angle estimation algorithms, such as multiple signal classification (MUSIC), estimation of signal parameters via rotational invariance technique (ESPRIT) and their variants have attracted enormous interests inside the array processing community due to their high resolution angle estimation [31]–[33] Their applications in massive MIMO systems and full-dimension MIMO systems for twodimensional angles estimation have been extensively studied in [34]–[38]. We further derive a simple expression for the theoretical bounds of mean squared errors (MSEs) performance in high signal-to-noise ratio (SNR) region, as well as the corresponding Cramér-Rao lower bounds (CRLBs) Both theoretical and numerical results are provided to corroborate the effectiveness of the proposed method. Notations: Small and upper bold-face letters donate column vectors and matrices, respectively; the superscripts (·)H , (·)T , (·)∗, (·)−1, (·)† stand for the conjugate-transpose, transpose, conjugate, inverse, pseudo-inverse of a matrix, respectively; tr(A) donates the trace of A; [A]ij is the (i, j)th entry of A; Diag{a} denotes a diagonal matrix with the diagonal element constructed from a, while Diag{A} denotes a vector whose elements are extracted from the diagonal components of A; vec(A) denotes the vectorization of A; R{A} denotes the real part of A; S{A} denotes the Imaginary part of A; [a]i:j denotes the subvector of a that starts with [a]i and ends at [a]j ; [A]i:j denotes the submatrix of A that starts with row [a]i,: and ends at row [a]j,:; E{·} denotes the statistical expectation, and h is the Euclidean norm of h
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
