arXiv subjects

Shun Zhang

Publications and source records attributed to Shun Zhang.

At least 91 records · Page 5Linked to original sources

arXiv2021

Deep Learning Based Antenna-time Domain Channel Extrapolation for Hybrid mmWave Massive MIMO

In a time-varying massive multiple-input multipleoutput (MIMO) system, the acquisition of the downlink channel state information at the base station (BS) is a very challenging task due to the prohibitively high overheads associated with downlink training and uplink feedback. In this paper, we consider the hybrid precoding structure at BS and examine the antennatime domain channel extrapolation. We design a latent ordinary differential equation (ODE)-based network under the variational auto-encoder (VAE) framework to learn the mapping function from the partial uplink channels to the full downlink ones at the BS side. Specifically, the gated recurrent unit is adopted for the encoder and the fully-connected neural network is used for the decoder. The end-to-end learning is utilized to optimize the network parameters. Simulation results show that the designed network can efficiently infer the full downlink channels from the partial uplink ones, which can significantly reduce the channel training overhead.

cs.IT↗

arXiv2021

Cascaded Channel Estimation for RIS Assisted mmWave MIMO Transmissions

Channel estimation is challenging for the reconfigurable intelligence surface (RIS) assisted millimeter wave (mmWave) communications. Since the number of coefficients of the cascaded channels in such systems is closely dependent on the product of the number of base station antennas and the number of RIS elements, the pilot overhead would be prohibitively high. In this letter, we propose a cascaded channel estimation framework for an RIS assisted mmWave multiple-input multiple-output system, where the wideband effect on transmission model is considered. Then, we transform the wideband channel estimation into a parameter recovery problem and use a few pilot symbols to detect the channel parameters by the Newtonized orthogonal matching pursuit algorithm. Moreover, the Cramer-Rao lower bound on the channel estimation is introduced. Numerical results show the effectiveness of the proposed channel estimation scheme.

eess.SP↗

arXiv2021

AIRIS: Artificial Intelligence Enhanced Signal Processing in Reconfigurable Intelligent Surface Communications

Reconfigurable intelligent surface (RIS) is an emerging meta-surface that can provide additional communications links through reflecting the signals, and has been recognized as a strong candidate of 6G mobile communications systems. Meanwhile, it has been recently admitted that implementing artificial intelligence (AI) into RIS communications will extensively benefit the reconfiguration capacity and enhance the robustness to complicated transmission environments. Besides the conventional model-driven approaches, AI can also deal with the existing signal processing problems in a data-driven manner via digging the inherent characteristic from the real data. Hence, AI is particularly suitable for the signal processing problems over RIS networks under unideal scenarios like modeling mismatching, insufficient resource, hardware impairment, as well as dynamical transmissions. As one of the earliest survey papers, we will introduce the merging of AI and RIS, called AIRIS, over various signal processing topics, including environmental sensing, channel acquisition, beamforming design, and resource scheduling, etc. We will also discuss the challenges of AIRIS and present some interesting future directions.

eess.SP↗

arXiv2021

Deep Learning Based RIS Channel Extrapolation with Element-grouping

Reconfigurable intelligent surface (RIS) is considered as a revolutionary technology for future wireless communication networks. In this letter, we consider the acquisition of the cascaded channels, which is a challenging task due to the massive number of passive RIS elements. To reduce the pilot overhead, we adopt the element-grouping strategy, where each element in one group shares the same reflection coefficient and is assumed to have the same channel condition. We analyze the channel interference caused by the element-grouping strategy and further design two deep learning based networks. The first one aims to refine the partial channels by eliminating the interference, while the second one tries to extrapolate the full channels from the refined partial channels. We cascade the two networks and jointly train them. Simulation results show that the proposed scheme provides significant gain compared to the conventional element-grouping method without interference elimination.

eess.SP↗

arXiv2021

Deep Learning based Channel Extrapolation for Large-Scale Antenna Systems: Opportunities, Challenges and Solutions

With the depletion of spectrum, wireless communication systems turn to exploit large antenna arrays to achieve the degree of freedom in space domain, such as millimeter wave massive multi-input multioutput (MIMO), reconfigurable intelligent surface assisted communications and cell-free massive MIMO. In these systems, how to acquire accurate channel state information (CSI) is difficult and becomes a bottleneck of the communication links. In this article, we introduce the concept of channel extrapolation that relies on a small portion of channel parameters to infer the remaining channel parameters. Since the substance of channel extrapolation is a mapping from one parameter subspace to another, we can resort to deep learning (DL), a powerful learning architecture, to approximate such mapping function. Specifically, we first analyze the requirements, conditions and challenges for channel extrapolation. Then, we present three typical extrapolations over the antenna dimension, the frequency dimension, and the physical terminal, respectively. We also illustrate their respective principles, design challenges and DL strategies. It will be seen that channel extrapolation could greatly reduce the transmission overhead and subsequently enhance the performance gains compared with the traditional strategies. In the end, we provide several potential research directions on channel extrapolation for future intelligent communications systems.

cs.IT↗

arXiv2021

Deep Learning based Antenna Selection and CSI Extrapolation in Massive MIMO Systems

A critical bottleneck of massive multiple-input multiple-output (MIMO) system is the huge training overhead caused by downlink transmission, like channel estimation, downlink beamforming and covariance observation. In this paper, we propose to use the channel state information (CSI) of a small number of antennas to extrapolate the CSI of the other antennas and reduce the training overhead. Specifically, we design a deep neural network that we call an antenna domain extrapolation network (ADEN) that can exploit the correlation function among antennas. We then propose a deep learning (DL) based antenna selection network (ASN) that can select a limited antennas for optimizing the extrapolation, which is conventionally a type of combinatorial optimization and is difficult to solve. We trickly designed a constrained degradation algorithm to generate a differentiable approximation of the discrete antenna selection vector such that the back-propagation of the neural network can be guaranteed. Numerical results show that the proposed ADEN outperforms the traditional fully connected one, and the antenna selection scheme learned by ASN is much better than the trivially used uniform selection.

eess.SP↗

arXiv2020

Ordinary Differential Equation-based CNN for Channel Extrapolation over RIS-assisted Communication

The reconfigurable intelligent surface (RIS) is considered as a promising new technology for reconfiguring wireless communication environments. To acquire the channel information accurately and efficiently, we only turn on a fraction of all the RIS elements, formulate a sub-sampled RIS channel, and design a deep learning based scheme to extrapolate the full channel information from the partial one. Specifically, inspired by the ordinary differential equation (ODE), we set up connections between different data layers in a convolutional neural network (CNN) and improve its structure. Simulation results are provided to demonstrate that our proposed ODE-based CNN structure can achieve faster convergence speed and better solution than the cascaded CNN.

cs.IT↗

arXiv2020

A Differentially Private Framework for Spatial Crowdsourcing with Historical Data Learning

Spatial crowdsourcing (SC) is an increasing popular category of crowdsourcing in the era of mobile Internet and sharing economy. It requires workers to arrive at a particular location for task fulfillment. Effective protection of location privacy is essential for workers' enthusiasm and valid task assignment. However, existing SC models with differential privacy usually perturb real-time location data for both partition and data publication. Such a way may produce large perturbations to counting queries that affect assignment success rate and allocation accuracy. This paper proposes a framework (R-HT) for protecting location privacy of workers taking advantage of both real-time and historical data. We simulate locations by sampling the probability distribution learned from historical data, use them for grid partition, and then publish real-time data under this partitioning with differential privacy. This realizes that most privacy budget is allocated to the worker count of each cell and yields an improved Private Spatial Decomposition approach. Moreover, we introduce some strategies for geocast region construction, including quality scoring function and local maximum geocast radius. A series of experimental results on real-world datasets shows that R-HT attains a stable success rate of task assignment, saves performance overhead and fits for dynamic assignment on crowdsourcing platforms.

cs.CR↗

arXiv2020

Utility-efficient Differentially Private K-means Clustering based on Cluster Merging

Differential privacy is widely used in data analysis. State-of-the-art $k$-means clustering algorithms with differential privacy typically add an equal amount of noise to centroids for each iterative computation. In this paper, we propose a novel differentially private $k$-means clustering algorithm, DP-KCCM, that significantly improves the utility of clustering by adding adaptive noise and merging clusters. Specifically, to obtain $k$ clusters with differential privacy, the algorithm first generates $n \times k$ initial centroids, adds adaptive noise for each iteration to get $n \times k$ clusters, and finally merges these clusters into $k$ ones. We theoretically prove the differential privacy of the proposed algorithm. Surprisingly, extensive experimental results show that: 1) cluster merging with equal amounts of noise improves the utility somewhat; 2) although adding adaptive noise only does not improve the utility, combining both cluster merging and adaptive noise further improves the utility significantly.

cs.CR↗

arXiv2020

Primal-Dual Reduced Basis Methods for Convex Minimization Variational Problems: Robust True Solution A Posteriori Error Certification and Adaptive Greedy Algorithms

In this paper, with the parametric symmetric coercive elliptic boundary value problem as an example of the primal-dual variational problems satisfying the strong duality, we develop primal-dual reduced basis methods (PD-RBM) with robust true error certifications and discuss three versions of greedy algorithms to balance the finite element error, the exact reduced basis error, and the adaptive mesh refinements. For a class of convex minimization variational problems which has corresponding dual problems satisfying the strong duality, the primal-dual gap between the primal and dual functionals can be used as a posteriori error estimator. This primal-dual gap error estimator is robust with respect to the parameters of the problem, and it can be used for both mesh refinements of finite element methods and the true RB error certification. With the help of integrations by parts formula, the primal-dual variational theory is developed for the symmetric coercive elliptic boundary value problems with non-homogeneous boundary conditions by both the conjugate function and Lagrangian theories. A generalized Prager-Synge identity, which is the primal-dual gap error representation for this specific problem, is developed. RBMs for both the primal and dual problems with robust error estimates are developed. The dual variational problem often can be viewed as a constraint optimization problem. In the paper, different from the standard saddle-point finite element approximation, the dual RBM is treated as a Galerkin projection by constructing RB spaces satisfying the homogeneous constraint. Inspired by the greedy algorithm with spatio-parameter adaptivity of \cite{Yano:18}, adaptive balanced greedy algorithms with primal-dual finite element and reduced basis error estimators are discussed. Numerical tests are presented to test the PD-RBM with adaptive balanced greedy algorithms.

math.NA↗

arXiv2020

Deep Learning Optimized Sparse Antenna Activation for Reconfigurable Intelligent Surface Assisted Communication

To capture the communications gain of the massive radiating elements with low power cost, the conventional reconfigurable intelligent surface (RIS) usually works in passive mode. However, due to the cascaded channel structure and the lack of signal processing ability, it is difficult for RIS to obtain the individual channel state information and optimize the beamforming vector. In this paper, we add signal processing units for a few antennas at RIS to partially acquire the channels. To solve the crucial active antenna selection problem, we construct an active antenna selection network that utilizes the probabilistic sampling theory to select the optimal locations of these active antennas. With this active antenna selection network, we further design two deep learning (DL) based schemes, i.e., the channel extrapolation scheme and the beam searching scheme, to enable the RIS communication system. The former utilizes the selection network and a convolutional neural network to extrapolate the full channels from the partial channels received by the active RIS antennas, while the latter adopts a fully-connected neural network to achieve the direct mapping between the partial channels and the optimal beamforming vector with maximal transmission rate. Simulation results are provided to demonstrate the effectiveness of the designed DL-based schemes.

eess.SP↗

arXiv2020

Deep Learning Based Antenna Selection for Channel Extrapolation in FDD Massive MIMO

In massive multiple-input multiple-output (MIMO) systems, the large number of antennas would bring a great challenge for the acquisition of the accurate channel state information, especially in the frequency division duplex mode. To overcome the bottleneck of the limited number of radio links in hybrid beamforming, we utilize the neural networks (NNs) to capture the inherent connection between the uplink and downlink channel data sets and extrapolate the downlink channels from a subset of the uplink channel state information. We study the antenna subset selection problem in order to achieve the best channel extrapolation and decrease the data size of NNs. The probabilistic sampling theory is utilized to approximate the discrete antenna selection as a continuous and differentiable function, which makes the back propagation of the deep learning feasible. Then, we design the proper off-line training strategy to optimize both the antenna selection pattern and the extrapolation NNs. Finally, numerical results are presented to verify the effectiveness of our proposed massive MIMO channel extrapolation algorithm.

eess.SP↗

arXiv2020

The Largest Entry in the Inverse of a Vandermonde Matrix

We investigate the size of the largest entry (in absolute value) in the inverse of certain Vandermonde matrices. More precisely, for every real $b > 1$, let $M_b(n)$ be the maximum of the absolute values of the entries of the inverse of the $n \times n$ matrix $[b^{i j}]_{0 \leq i, j < n}$. We prove that $\lim_{n \to +\infty} M_b(n)$ exists, and we provide some formulas for it.

math.RA↗

arXiv2020

Graph Neural Network based Channel Tracking for Massive MIMO Networks

In this paper, we resort to the graph neural network (GNN) and propose the new channel tracking method for the massive multiple-input multiple-output networks under the high mobility scenario. We first utilize a small number of pilots to achieve the initial channel estimation. Then, we represent the obtained channel data in the form of graphs and describe the channel spatial correlation by the weights along the edges of the graph. Furthermore, we introduce the computation steps of the main unit for the GNN and design a GNN-based channel tracking framework, which includes an encoder, a core network and a decoder. Simulation results corroborate that our proposed GNN-based scheme can achieve better performance than the works with feedforward neural network.

cs.IT↗

arXiv2020

A New Path Division Multiple Access for the Massive MIMO-OTFS Networks

This paper focuses on a new path division multiple access (PDMA) for both uplink (UL) and downlink (DL) massive multiple-input multiple-output network over a high mobility scenario, where the orthogonal time frequency space (OTFS) is adopted. First, the 3D UL channel model and the received signal model in the angle-delay-Doppler domain are studied. Secondly, the 3D-Newtonized orthogonal matching pursuit algorithm is utilized for the extraction of the UL channel parameters, including channel gains, directions of arrival, delays, and Doppler frequencies, over the antenna-time-frequency domain. Thirdly, we carefully analyze energy dispersion and power leakage of the 3D angle-delay-Doppler channels. Then, along UL, we design a path scheduling algorithm to properly assign angle-domain resources at user sides and to assure that the observation regions for different users do not overlap over the 3D cubic area, i.e., angle-delay-Doppler domain. After scheduling, different users can map their respective data to the scheduled delay-Doppler domain grids, and simultaneously send the data to base station (BS) without inter-user interference in the same OTFS block. Correspondingly, the signals at desired grids within the 3D resource space of BS are separately collected to implement the 3D channel estimation and maximal ratio combining-based data detection over the angle-delay-Doppler domain. Then, we construct a low complexity beamforming scheme over the angle-delay-Domain domain to achieve inter-user interference free DL communication. Simulation results are provided to demonstrate the validity of our proposed unified UL/DL PDMA scheme.

eess.SP↗

arXiv2020

Uplink-aided High Mobility Downlink Channel Estimation over Massive MIMO-OTFS System

Although it is often used in the orthogonal frequency division multiplexing (OFDM) systems, application of massive multiple-input multiple-output (MIMO) over the orthogonal time frequency space (OTFS) modulation could suffer from enormous training overhead in high mobility scenarios. In this paper, we propose one uplink-aided high mobility downlink channel estimation scheme for the massive MIMO-OTFS networks. Specifically, we firstly formulate the time domain massive MIMO-OTFS signal model along the uplink and adopt the expectation maximization based variational Bayesian (EM-VB) framework to recover the uplink channel parameters including the angle, the delay, the Doppler frequency, and the channel gain for each physical scattering path. Correspondingly, with the help of the fast Bayesian inference, one low complex approach is constructed to overcome the bottleneck of the EM-VB. Then, we fully exploit the angle, delay and Doppler reciprocity between the uplink and the downlink and reconstruct the angles, the delays, and the Doppler frequencies for the downlink massive channels at the base station. Furthermore, we examine the downlink massive MIMO channel estimation over the delay-Doppler-angle domain. The channel dispersion of the OTFS over the delay-Doppler domain is carefully analyzed. Various numerical examples are presented to confirm the validity and robustness of the proposed scheme.

eess.SP↗

arXiv2020

RSANet: Recurrent Slice-wise Attention Network for Multiple Sclerosis Lesion Segmentation

Brain lesion volume measured on T2 weighted MRI images is a clinically important disease marker in multiple sclerosis (MS). Manual delineation of MS lesions is a time-consuming and highly operator-dependent task, which is influenced by lesion size, shape and conspicuity. Recently, automated lesion segmentation algorithms based on deep neural networks have been developed with promising results. In this paper, we propose a novel recurrent slice-wise attention network (RSANet), which models 3D MRI images as sequences of slices and captures long-range dependencies through a recurrent manner to utilize contextual information of MS lesions. Experiments on a dataset with 43 patients show that the proposed method outperforms the state-of-the-art approaches. Our implementation is available online at https://github.com/tinymilky/RSANet.

eess.IV↗

arXiv2020

Generalized Prager-Synge Inequality and Equilibrated Error Estimators for Discontinuous Elements

The well-known Prager-Synge identity is valid in $H^1(Ω)$ and serves as a foundation for developing equilibrated a posteriori error estimators for continuous elements. In this paper, we introduce a new inequality, that may be regarded as a generalization of the Prager-Synge identity, to be valid for piecewise $H^1(Ω)$ functions for diffusion problems. The inequality is proved to be identity in two dimensions. For nonconforming finite element approximation of arbitrary odd order, we propose a fully explicit approach that recovers an equilibrated flux in $H(div; Ω)$ through a local element-wise scheme and that recovers a gradient in $H(curl;Ω)$ through a simple averaging technique over edges. The resulting error estimator is then proved to be globally reliable and locally efficient. Moreover, the reliability and efficiency constants are independent of the jump of the diffusion coefficient regardless of its distribution.

math.NA↗