Covariance Recovery and CRB Analysis for DOA Estimation by Mixed One-Bit Quantized Signals

To realize the direction-of-arrivals (DoAs) estimation without the prior information about the multipath number, a novel method using the deep learning is introduced for the millimeter (mmWave) massive multiple-input and multiple-output (MIMO) systems. In particular, the DoAs estimation is decomposed into three sub-problems, which are solved by the corresponding convolutional neural network (CNN), respectively. The estimation of multipath number is achieved by a multi-label classification model using the proposed CNN-I. Then, the proposed CNN-II is introduced for the DoA estimation of the line-of-sight (LOS) path. Based on the predicted multipath number obtained by the CNN-I, a regression model using the proposed CNN-III is applied for the DoAs estimation of non-line-of-sight (NLOS) paths. The proposed DoAs estimation method is implemented by learning the non-linear relationship between the sample covariance matrix of the received signal and the angles, thus reconstructing the mapping model. A series of results validate the superiority of the proposed DoAs estimation method in low signal-to-noise-ratio (SNR) regimes. The CNN-I is capable of achieving the good multipath number estimation performance across a range of SNRs. The classification model based on the proposed CNN-II and the regression model using the proposed CNN-III achieve the lower DoAs estimation root mean square error (RMSE) compared with some deep learning-based methods, estimation of signal parameters via rotational invariance techniques (ESPRIT) and Root-MUltiple SIgnal Classification (Root-MUSIC) methods. Remarkably, the proposed method using the CNN-II exhibits the good robustness in the DoA estimation of the LOS path. Furthermore, the low DoAs estimation RMSE is also achieved in the uniform planar array.

Many research efforts were recently spent to address the challenging problem of DOA and angular spread estimation. However, to the best of our knowledge, no previous work has thoroughly investigated their specific impact on the performance of DOA-based antenna-array beamforming. In this contribution, we address this issue in the particular context of wideband CDMA using pilot-assisted or blind antenna-array receivers. In the process, we also assess whether the generalized channel-matched beamforming (i.e., without DOA estimation or a priori knowledge of the spatial structure of the channel) offers a better alternative. Link-level simulation results in terms of required SNR at target BER of 1% suggest that wideband CDMA array-receivers, whether pilot-assisted or blind, are extremely sensitive to angular spread mismatches and that the benefits of exploiting the spatial structure of the channel (i.e., estimation of DOA, angular spread, etc..) translate, at best, in negligible SNR gains when accurate channel identification is already implemented. These results call for implementing channel-matched instead of conventional DOA-based beamforming.

This study proposes a new uniform concentric circular (UCCA) shape of electronically steerable parasitic array radiator (ESPAR) antennas for directions of arrival (DoAs) estimation problem. The well-known estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithm is adapted to this special shape of UCCA and the authors demonstrate that the resulting algorithm yields better DoAs estimation accuracy and solves the failure of estimation problem when the signals' DoAs are in some particular sectors: notably for [0°–30°] and [160°–180°]. This approach has shown interesting performances since it can ensure good DoAs estimation over the entire azimuth plane with a low extra computational overhead. The constraints are the same imposed to the ESPRIT algorithm allowing about 66% reduction on the required computational efforts. The Cramer Rao bound on the variance of DoAs estimated by the proposed array geometry is analysed. Through simulation results, the authors demonstrate that applying ESPRIT in conjunction with the proposed antennas shape not only provides superior high-resolution localisation capabilities but also it drastically reduces computations compared with previous works.

Target detection is critical in many mission critical sensors and sensor network (MC-SSN) applications. For target detection in complicated electromagnetic environment, DOA estimation using polarization sensitive array (PSA) has been receiving increased attentions. In this paper, we propose the parallel co-prime polarization sensitive array (PCP-PSA) which consists of the cocentered orthogonal dipole triads (CODTs) to estimate two-dimensional direction-of-arrival (2D DOA) and polarization parameters. The degrees of freedom (DOFs) have been extended due to the co-prime structure, so that the more signals can be detected and the estimation accuracy is improved. In order to reduce the computation complexity, we construct a new cross-covariance matrix based on the CODTs, which converts the two-dimensional DOA estimation into two independent one-dimensional DOA estimations. Then, the spatial smoothing-based multiple signal classification algorithm(MUSIC) and the sparse representation-based method are applied to estimate 2D DOA with only one-dimensional (1D) peak searching and 1D dictionary, respectively. Finally, the polarization parameters are estimated by using the cross-covariance matrix between components of electric field vector. Compared with previous PSA-based algorithms, the proposed algorithm based on PCP-PSA can solve the underdetermined 2D DOA and polarization parameters estimation problem and has better estimation accuracy. Theoretical analyses and simulation results verify the effectiveness of the proposed methods in terms of computational complexity and estimation accuracy.

In this paper, a new design of Electronically Steerable Parasitic Array Radiator (ESPAR) antennas is proposed for directions of arrival (DoAs) estimation of coherent sources by the well known ESPRIT algorithm. The Spatial Smoothing technique (SS) is adapted to this antennas shape to remove signals' coherence. The excellent performances of the uniform circular geometry are used to ensure accurate DoAs estimation in all the azimuth coverage. Simulation results show that the use of the spatial smoothing technique with the proposed ESPAR antenna shape improves DoAs estimation accuracy compared to the original ESPRIT algorithm for ESPAR antennas especially in highly correlated sources situations and prove the validity of our approach.

Directions of arrival (DoAs) estimation of multiple sources using an antenna array is a challenging topic in wireless communication. The DoAs estimation accuracy depends not only on the selected technique and algorithm, but also on the geometrical configuration of the antenna array used during the estimation. In this article the robustness of common planar antenna arrays against unaccounted mutual coupling is examined and their DoAs estimation capabilities are compared and analysed through computer simulations using the well-known MUltiple SIgnal Classification (MUSIC) algorithm. Our analysis is based on an electromagnetic concept to calculate an approximation of the impedance matrices that define the mutual coupling matrix (MCM). Furthermore, a CRB analysis is presented and used as an asymptotic performance benchmark of the studied antenna arrays. The impact of the studied antenna arrays geometry on the MCM structure is also investigated. Simulation results show that the UCCA has more robustness against unaccounted mutual coupling and performs better results than both UCA and URA geometries. The performed simulations confirm also that, although the UCCA achieves better performance under complicated scenarios, the URA shows better asymptotic (CRB) behaviour which promises more accuracy on DoAs estimation.

ABSTRACTBased on an extended reactance domain (RD) covariance matrix, this article proposes new alternatives for directions of arrival (DoAs) estimation of narrowband sources through an electronically steerable parasitic array radiator (ESPAR) antennas. Because of the centro symmetry of the classic ESPAR antennas, an unitary transformation is applied to the collected data that allow an important reduction in both computational cost and processing time and, also, an enhancement of the resolution capabilities of the proposed algorithms. Moreover, this article proposes a new approach for eigenvalues estimation through only some linear operations. The developed DoAs estimation algorithms based on this new approach has illustrated a good behaviour with less calculation cost and processing time as compared to other schemes based on the classic eigenvalues approach. The conducted simulations demonstrate that high-precision and high-resolution DoAs estimation can be reached especially in very closely sources situation and low sources power as compared to the RD-MUSIC algorithm and the RD-PM algorithm. The asymptotic behaviours of the proposed DoAs estimators are analysed in various scenarios and compared with the Cramer-Rao bound (CRB). The conducted simulations testify the high-resolution of the developed algorithms and prove the efficiently of the proposed approach.

DOA estimation is an important research area in array signal processing. Bayesian maximum a posterior probability density DOA estimator (BM DOA estimator) has been shown to perform excellently. However, the BM estimator requires a multidimensional grid search and the computational burden increases exponentially with the dimension. So it is difficult to be used in realtime applications. In order to reduce the computation, Monte Carlo methods are combined with BM DOA estimator. A novel Bayesian maximum a posterior DOA estimator based on importance sampling (ISBM DOA estimator) is proposed in this paper. ISBM DOA estimator not only keeps the good performance of the original BM DOA estimator, but also reduces the computation obviously because it needs not multidimensional search and reduces the computational complexity of the original method from O(L/sup K/) to O(K/spl times/H). Simulation results show that ISBM DOA estimator keeps the excellent performance of BM DOA estimator, but also reduces the computation evidently and performs better than MLE, MUSIC and MiniNorm, especially in low SNRs.

A method for full-azimuth DoA estimation of multiple signals with a hexagonal array is proposed. The DoA estimation is performed in two steps. In the first, a set of estimate candidates is constructed by gathering the estimates that are obtained from applying the Unitary-ESPRIT algorithm to several translational invariances designed into a hexagonal array. In the second step, the DoA estimates are successively selected from the estimate candidate set by using a selection function. The proposed method removes the north-or-south signal membership ambiguity and the limitation on the number of estimable sources, problems common to any ESPRIT-based algorithm used with one translational invariance. Therefore, up to M - 1 signal DoA estimations can be expected with an M-element hexagonal array in the full azimuth. The successive-selection approach is based on a selection function that uses an estimate of the signal's spatial correlation matrix to successively select the DoA estimates. For each DoA estimate selection, the already estimated signal components are removed from the correlation matrix. The method's DoA estimation and resolution capabilities are demonstrated by computer simulation.

The performance of DOA estimation with scalar sensor arrays using spatial sparse signal reconstruction (SSR) technique is affected by the grid spacing. In this paper, we formulate the DOA estimation with the acoustic vector sensor (AVS) arrays under SSR framework. A coarse-to-fine DOA estimation algorithm has been developed. The source spatial sparsity and the inter-relations among the manifold matrices of the AVS subarrays are jointly utilized to eliminate the grid effect in the SSR technique and the improvement of the overall DOA estimation performance is achieved at low complexity. Simulation results show that the proposed method effectively mitigates the DOA estimation bias caused by off-grid sources. Interestingly, our method gives good DOA estimation accuracy when sources are closely located.

The massive multiple input multiple output (MIMO) system is a prodigious method that could improvise a communication system's capacity with the aid of an increased number of antennas. The fundamental challenges of massive MIMO system are complicated channel modeling, high dimensional channel state information (CSI), and narrow radio frequency chains (RF chains), etc. For solving this problem we introduce deep neural network (DNN) framework for direction of arrival estimation (DOA estimation). Using the information obtained about DOA and gain that is complex, the channel is estimated and mean square error (MSE) performance is evaluated for both DOA and channel estimation.

The positioning ability of the DOA estimation algorithms in array signal processing depends on the correct selection of snapshot data including the target echo signal. However, in the scene with low signal-to-noise ratio (SNR), the CFAR method, which is frequently used before DOA estimation, cannot effectively discover the snapshot data containing the targets, resulting in the loss of the target. In this paper, we propose a multi-frame DOA (MF-DOA) estimation algorithm, which implements multi-frame energy accumulation and snapshot data extraction based on the introduced target motion model so that the weak moving targets can be effectively located. The simulation data and real MIMO radar data are processed, and the results show that the MF-DOA algorithm has good ability to excavate weak moving targets and extend the radar detection range.

A dual parallel factor (PARAFAC)-based approach to jointly estimating the two-dimensional direction of arrival (2-D DOA) and Doppler is proposed in this paper, where an L-shaped array consisting of acoustic vector-sensor is used. First, we apply the PARAFAC decomposition to the data model formed by concatenating the outputs of multi-level delays of the observations, and we get the parameter matrix H, which accomplish the 2-D DOA estimation and pairing automatically, then the dual PARAFAC decomposition is applied to the achieved composite steering matrix from the first PARAFAC decomposition, and thus, the same permutation matrices link the estimates of steering matrices and delay matrices from X-subarray and Y-subarray, respectively. Following this, the Doppler and 2-D DOA matching information are obtained via triple matching implementation, e.g. 2-D DOA and frequency matching. Finally, Doppler is estimated by delay matrices. The proposed algorithm is computationally effective for both uniform and non-uniform L-shaped array as SNR exceeds 15dB, and its performance outperforms the joint angle and Doppler shift ESPRIT (JAD-ESPRIT) algorithm and the joint angle and Doppler shift PM (JAD-PM) algorithm. The simulation results justified the effectiveness of the proposed algorithm.

Signal processing in spherical harmonic domain has the ability to decouple frequency dependent and location dependent components of the signal received. A method for low dimensional spherical harmonic feature extraction is proposed in this work for DOA estimation in noisy and reverberant environments. The features are extracted using frequency smoothing and a transformation which makes them frequency and signal invariant. Additionally an online manifold regularization framework is explored which utilizes the proposed spherical harmonic features to compute real time DOA estimates. This framework minimizes an instantaneous risk function and finds an inverse mapping function that maps spherical harmonic features to the DOA estimate. Performance of the proposed DOA estimation method is then compared with DOA estimates obtained from features such as generalized cross correlation and relative transfer function in a semi-supervised manifold regularization framework. Experimental results on DOA estimation in terms of root mean square error and probability of resolution indicate a reasonable improvement in the localization performance along with significant reduction in feature dimension.

Sparse linear arrays (SLAs), such as nested and co-prime arrays, have the attractive capability of providing enhanced degrees of freedom by exploiting the co-array model. Accordingly, co-array-based Direction of Arrivals (Doas) estimation has recently gained considerable interest in array processing. The literature has suggested applying MUSIC on an augmented sample covariance matrix for co-array-based Doas estimation. In this paper, we propose a Least Squares (LS) estimator for co-array-based DoAs estimation employing the covariance fitting method as an alternative to MUSIC. We show that the proposed LS estimator provides consistent estimates of Doas of identifiable sources for SLAs. Additionally, an analytical expression for the large sample performance of the proposed estimator is derived. Numerical results illustrate the finite sample behavior in relation to the derived analytical expression. Moreover, the performance of the proposed LS estimator is compared to the co-array-based MUSIC.