Dynamic Compressive Sensing Research Articles

Temporally-correlated massive access: joint user activity detection and channel estimation via vector approximate message passing

In the paper, we investigate a massive machine-type communication (mMTC), where numerous single-antenna users communicate with a single-antenna base station while being active. However, the status of user can undergoes multiple transitions between active and inactive states across whole consecutive intervals. Then, we formulate the problem of joint user activity detection and channel estimation within the dynamic compressed sensing (DCS) framework, considering the temporally-correlated user activity across the entire consecutive intervals. To be specific, we introduce a new hybrid vector approximate message passing algorithm for DCS (HyVAMP-DCS). The proposed algorithm comprises a VAMP block for estimating channel and a loopy belief propagation (LBP) block for detecting user activity. Moreover, these two blocks can exchange messages, enhancing the performance of both channel estimation and user activity detection. Importantly, compared to the fragile GAMP algorithm, VAMP is robust and applicable to a much broader class of large random matrices. Furthermore, the fixed points of VAMP’s state evolution align with the replica prediction of the minimum mean-squared error. The simulation results illustrate the superiority of HyVAMP-DCS, demonstrating its significant outperformance over HyGAMP-DCS.

Dynamic Compressed Sensing Approach for Unsourced Random Access

Dynamic Compressed Sensing Approach for Unsourced Random Access

Message Passing-Based Joint User Activity Detection and Channel Estimation for Temporally-Correlated Massive Access

This paper studies the user activity detection and channel estimation problem in a temporally-correlated massive access system where a very large number of users communicate with a base station sporadically and each user once activated can transmit with a large probability over multiple consecutive frames. We formulate the problem as a dynamic compressed sensing (DCS) problem to exploit both the sparsity and the temporal correlation of user activity. By leveraging the hybrid generalized approximate message passing (HyGAMP) framework, we design a computationally efficient algorithm, HyGAMP-DCS, to solve this problem. In contrast to only exploiting the historical estimations, the proposed algorithm performs bidirectional message passing between the neighboring frames for activity likelihood update to fully exploit the temporally-correlated user activities. Furthermore, we develop an expectation maximization HyGAMP-DCS (EM-HyGAMP-DCS) algorithm to adaptively learn the hyperparameters during the estimation procedure when the system statistics are unknown. In particular, we propose to utilize the analysis tool of state evolution to find the appropriate hyperparameter initialization of EM-HyGAMP-DCS. Simulation results demonstrate that our proposed algorithms can significantly improve the user activity detection accuracy and reduce the channel estimation error.

Microvibration streaming measurements using dynamic compressed sensing for satellites

Long-term monitoring of satellite microvibrations generates a significant amount of data streams, placing strain on satellites with limited transmission capacity. To relieve this transmission strain, a dynamic compressed sensing (CS) framework is proposed for measuring satellite microvibrations. Microvibration streams are measured block by block and then reconstructed using a dynamic recovery algorithm. The recovery solution of one block can be used as a priori knowledge for the next block, allowing for faster updates. However, existing dynamic recovery algorithms are only applicable in the real domain and cannot be applied to microvibrations projected on a Fourier basis in the complex domain. In light of this event, the dynamic homotopy algorithm is expanded to the complex domain to deal with microvibration signals that are sparse in the Fourier basis. In comparison to conventional uniform sampling methods, the experimental results show that the dynamic CS with the expanded recovery algorithm can achieve a maximum root-mean-square acceleration (Grms) deviation of 4% in power spectrum density with one-fifth of the sampling points. Compared to recovery algorithms applicable to fixed measurements, the dynamic algorithm can achieve comparable accuracy in about one-third of the computation time. The experimental findings demonstrate the feasibility of satellite microvibrations measurements using dynamic CS.

Adaptive equalization based on dynamic compressive sensing for single-carrier multiple-input multiple-output underwater acoustic communications.

The sparse property of a direct adaptive equalizer (DAE) for single-carrier underwater acoustic communications is well recognized. It has been used to improve the performance and/or reduce the complexity of a DAE. Extensive investigations have been performed in terms of performance improvement. On the contrary, research on complexity reduction remains preliminary. A fundamental way for reducing the complexity of a DAE is to keep only significant taps while discarding trivial taps, that is, to run a partial-tap DAE. Existing partial-tap DAE designs assume a slowly varying sparse structure and may suffer performance degradation under a severe underwater environment. Motivated by this fact, the dynamic compressed sensing (DCS) technique is resorted to and a partial-tap DAE based on the sparse adaptive orthogonal matching pursuit-affine projection algorithm is proposed. The sparse adaptive orthogonal matching pursuit-affine projection algorithm-direct adaptive equalizer (SpAdOMP-APA-DAE) achieves symbol-wise updating of both positions and values of the significant coefficients. In this paper, a more extensive study on DCS-based DAEs is performed, and an enhanced dynamic compressed sensing-direct adaptive equalizer design enabled by the sparse adaptive subspace pursuit-improved proportionate affine projection algorithm (SpAdOMP-IPAPA) is proposed. The sparse adaptive subspace pursuit-improved proportionate affine projection algorithm-direct adaptive equalizer enjoys lower complexity while better performance than the previous SpAdOMP-APA-DAE. Experimental results corroborated the superiority of the SpAdOMP-IPAPA-DAE.

Performance Analysis of 5G Cooperative-NOMA for IoT-Intermittent Communication

Non-orthogonal multiple Access (NOMA) is a potential 5G era multiple-access scheme that is proposed for the future mobile Internet and IoT applications which will require enormous increase in data traffic, massive-number of devices connectivity, high spectral efficiency, low-overhead and low-latency. It utilizes the same time-slots, frequency and spreading-codes for all the users. It uses the power-domain and assign different power levels to users for multiple access. The uplink (UL) communication in the present 4G-Networks is performed by the base station (BS) that uses a request-grant mechanism in which a large-overhead and latency is produced. This issue will get more severe in upcoming 5G-Networks. For this purpose, a grant-free NOMA for UL communication, in which dynamic compressed-sensing (DCS) algorithm will perform multi-user detection (MUD) as well as data-detection is proposed. It deploys the temporal-correlation of active-user sets (AUS) in adjacent time-slots from which the estimated AUS is used as the prior-knowledge to estimate AUS in the next time-slot. For the downlink (DL) communication, the proposed system performance evaluation is performed using Rician fading-channels for Cooperative Relaying System (CRS) NOMA. The simulations results show that the proposed DCS-MUD and CRS NOMA over Rician fading-channels perform much better than the conventional CS-MUD and traditional-CRS.

A Block Sparse-Based Dynamic Compressed Sensing Channel Estimator for Underwater Acoustic Communication

Due to the complex ocean propagation environments, the underwater acoustic (UWA) multipath channel often exhibits block sparse time-varying features, and while dynamic compressed sensing (DCS) can mitigate the time-varying effects of the UWA channel, DCS-based algorithms have limited performance for the UWA channel with block sparsity. In this study, by formulating the UWA channel with blocks concatenation, a block sparse-based DCS approach (BS-CS) is proposed to explore the block and time-varying sparsity of UWA channel simultaneously. In detail, we firstly adopt a block sparse recovery algorithm, block orthogonal matching pursuit (BOMP), to compute the temporary estimate. Then, the CS approach is applied to compute the support additions, which are caused by the time-varying components of the UWA channel. Next, we use the selected support to perform the BOMP estimate, and obtain the estimated channel response. Finally, the numerical simulation and the sea experiment were carried out to verify the superior performance of the proposed BS-CS algorithm in the block sparse time-varying UWA channels.

Exploiting Sparsity for Underwater Acoustic Sensor Network Under Time-Varying Channels

As a time-frequency doubly selective channel, severe multipath, Doppler, as well as the large time-delay characteristics of the underwater acoustic (UWA) channel pose significant challenge to the research and design of UWA communication and network systems. To mitigate these negative factors, inherent sparsity contained in the UWA channel has been extensively investigated to improve UWA communication via a sparsity exploitation receiver. While the performance of the UWA sensor network is highly dependent on that of the physical layer, there are few investigations reported on exploiting channel sparsity from the viewpoint of UWA networking. In this article, a UWA sensor network adopting the sparsity exploitation physical layer is evaluated based on the network simulator 3 (NS-3) simulation tool. The simulation time-varying channel is generated by incorporating the Bellhop channel model with the statistical characteristics extracted from experimental shallow water channels. Three types of physical layers, i.e., those do not adopt sparse exploitation, adopting compressed sensing (CS), as well as the dynamic CS (DCS) technique, are employed for evaluation and comparison of network behavior under different media access control (MAC) protocols. The evaluation results verify the effectiveness of sparsity exploitation in improving UWA sensor network performance in the presence of time variations, while giving a quantitative comparison between enhancement achieved by the CS and DCS sparsity exploitation.

ECG Monitoring Based on Dynamic Compressed Sensing of Multi-Lead Signals.

This paper presents an innovative method for multiple lead electrocardiogram (ECG) monitoring based on Compressed Sensing (CS). The proposed method extends to multiple leads signals, a dynamic Compressed Sensing method, that were previously developed on a single lead. The dynamic sensing method makes use of a sensing matrix in which its elements are dynamically obtained from the signal to be compressed. In this method, for the application to multiple leads, it is proposed to use a single sensing matrix for which its elements are obtained from a combination of multiple leads. The proposed method is evaluated on a wide set of signals and acquired on healthy subjects and on subjects affected by different pathologies, such as myocardial infarction, cardiomyopathy, and bundle branch block. The experimental results demonstrated that the proposed method can be adopted for a Compression Ratio () up to 10, without compromising signal quality. In particular, for 10, it exhibits a percentage of root-mean-squared difference average among a wide set of ECG signals lower than 3%.

Estimation of Underwater Acoustic Channel With Hybrid Sparsity via Static-Dynamic Discriminative Compressed Sensing

The technology of underwater acoustic communication plays an important role in linking the underwater sensors for data collection and networking. Due to simultaneous presence of direct path as well as reflections upon static or dynamic boundaries, underwater acoustic (UWA) channel exhibits hybrid sparse, i.e., consists of relatively static multipath arrivals and time varying ones. While the classic sparse channel estimation algorithms generally do not take this type of hybrid sparsity into account, dynamic compressed sensing (DCS) technique enables the exploitation of time varying sparsity by incorporating CS with Kalman filtering, namely, Kalman Filtered Compressed Sensing (KF-CS). However, by assuming the whole UWA channel as sparse set with time varying support, static sparse components will unavoidably lead to model mismatch and performance degradation. In this paper, by formulating the hybrid multipath UWA channels as sparse set consisting of static and time varying supports, a static-dynamic discriminative compressed sensing (SDD-CS) approach is proposed to explore the hybrid sparsity. Hybrid multipath arrivals are firstly decomposed into static and time varying components by discriminative orthogonal matching pursuit (DOMP), and then the magnitude of time varying and static multipath is estimated by KF-CS and SOMP respectively, after that estimated results of static and time varying component are summed up to obtain the whole channel response. Numerical simulations verify the superiority of the proposed SDD-CS algorithm. Finally experimental results obtained from a shallow water acoustic communication scenario are provided to demonstrate the effectiveness of the proposed algorithm, compared to conventional algorithms.

Exploiting dynamic sparsity for time reversal underwater acoustic communication under rapidly time varying channels

Underwater acoustic (UWA) communication encounters significant difficulties posed by the simultaneous presence of multipath and time variations. It has been recognized that, under the assumption that channel keeps static within the processing windows, time reversal (TR) processing approach is capable of mitigating multipath via the spatial–temporal focusing. However, for a rapidly time-varying UWA channel, the replica channel impulse response in time reversal processor need to be updated frequently to avoid performance degradation, which leads to huge computational complexity and significant overhead. Consider that after time reversal processing the resulting time-reversed channel response exhibits much more slow time variations compared to the original channel impulse response, previously Kalman filtering has been adopted to track the time-reversed channel response to alleviate the requirement of frequent channel update. Under the framework of dynamic compressed sensing (DCS), in this paper the time-reversed channel is formulated as a sparse set consisting of constant and slowly time-varying supports to derive the Kalman Filtered Compressed Sensing (KF-CS) channel estimation algorithm for time reversal communication. Based on the receiver structure by coupling the KF-CS estimator driven time reversal processor with a single channel decision-feedback equalizer (TR-DCS-DFE), shallow water experimental results with field data are provided to demonstrate the effectiveness of the proposed algorithm, compared to the classic time reversal communication scheme.

Hybrid Particle Filter Based Dynamic Compressed Sensing for Signal-Level Multitarget Tracking

We propose a novel algorithm, state propagation based dynamic compressed sensing (SP-DCS), that uses a target dynamic model in dynamic compressed sensing (DCS) to track a fixed number of targets. To track a time-varying number of targets using raw measurements from a Doppler radar, we also propose a novel hybrid particle filter based dynamic compressed sensing (HPF-DCS) algorithm. We calculate the support set in a Bayesian framework and a particle filter approximates the posterior probability mass function (pmf) of the support set. HPF-DCS is a combination of random and deterministic sampling. In random sampling, a number of predicted existing sub-particles are sampled from the prior pmf of the existing support set to handle the scenario when targets disappear randomly at a scan time. In deterministic sampling, the new support set corresponding to newly appearing targets is calculated by solving a sparsity promoting optimization problem. Our simulation results show that the proposed algorithm can track a time-varying number of targets successfully. It also outperforms the sequential Monte Carlo based probability hypothesis density (SMC-PHD) filter, as well as the multi-mode, multi-target track before detect (MM-MT-TBD) filter.

Performance Analysis of 5G Cooperative-NOMA for IoT-Intermittent Communication

Non-orthogonal multiple Access (NOMA) is a potential 5G era multiple-access scheme that is proposed for the future mobile Internet and IoT applications which will require enormous increase in data traffic, massive-number of devices connectivity, high spectral efficiency, low-overhead and low-latency. It utilizes the same time-slots, frequency and spreading-codes for all the users. It uses the power-domain and assign different power levels to users for multiple access. The uplink (UL) communication in the present 4G-Networks is performed by the base station (BS) that uses a request-grant mechanism in which a large-overhead and latency is produced. This issue will get more severe in upcoming 5G-Networks. For this purpose, a grant-free NOMA for UL communication, in which dynamic compressed-sensing (DCS) algorithm will perform multi-user detection (MUD) as well as data-detection is proposed. It deploys the temporal-correlation of active-user sets (AUS) in adjacent time-slots from which the estimated AUS is used as the prior-knowledge to estimate AUS in the next time-slot. For the downlink (DL) communication, the proposed system performance evaluation is performed using Rician fading-channels for Cooperative Relaying System (CRS) NOMA. The simulations results show that the proposed DCS-MUD and CRS NOMA over Rician fading-channels perform much better than the conventional CS-MUD and traditional-CRS.

Dynamic Compressive Sensing-Based Multi-User Detection for Uplink Grant-Free NOMA

Non-orthogonal multiple access (NOMA) can support more users than OMA techniques using the same wireless resources, which is expected to support massive connectivity for Internet of Things in 5G. Furthermore, in order to reduce the transmission latency and signaling overhead, grant-free transmission is highly expected in the uplink NOMA systems, where user activity has to be detected. In this letter, by exploiting the temporal correlation of active user sets, we propose a dynamic compressive sensing (DCS)-based multi-user detection (MUD) to realize both user activity and data detection in several continuous time slots. In particular, as the temporal correlation of the active user sets between adjacent time slots exists, we can use the estimated active user set in the current time slot as the prior information to estimate the active user set in the next time slot. Simulation results show that the proposed DCS-based MUD can achieve much better performance than that of the conventional CS-based MUD in NOMA systems.

Abdominal 4D flow MR imaging in a breath hold: combination of spiral sampling and dynamic compressed sensing for highly accelerated acquisition.

To develop a highly accelerated phase-contrast cardiac-gated volume flow measurement (four-dimensional [4D] flow) magnetic resonance (MR) imaging technique based on spiral sampling and dynamic compressed sensing and to compare this technique with established phase-contrast imaging techniques for the quantification of blood flow in abdominal vessels. This single-center prospective study was compliant with HIPAA and approved by the institutional review board. Ten subjects (nine men, one woman; mean age, 51 years; age range, 30-70 years) were enrolled. Seven patients had liver disease. Written informed consent was obtained from all participants. Two 4D flow acquisitions were performed in each subject, one with use of Cartesian sampling with respiratory tracking and the other with use of spiral sampling and a breath hold. Cartesian two-dimensional (2D) cine phase-contrast images were also acquired in the portal vein. Two observers independently assessed vessel conspicuity on phase-contrast three-dimensional angiograms. Quantitative flow parameters were measured by two independent observers in major abdominal vessels. Intertechnique concordance was quantified by using Bland-Altman and logistic regression analyses. There was moderate to substantial agreement in vessel conspicuity between 4D flow acquisitions in arteries and veins (κ = 0.71 and 0.61, respectively, for observer 1; κ = 0.71 and 0.44 for observer 2), whereas more artifacts were observed with spiral 4D flow (κ = 0.30 and 0.20). Quantitative measurements in abdominal vessels showed good equivalence between spiral and Cartesian 4D flow techniques (lower bound of the 95% confidence interval: 63%, 77%, 60%, and 64% for flow, area, average velocity, and peak velocity, respectively). For portal venous flow, spiral 4D flow was in better agreement with 2D cine phase-contrast flow (95% limits of agreement: -8.8 and 9.3 mL/sec, respectively) than was Cartesian 4D flow (95% limits of agreement: -10.6 and 14.6 mL/sec). The combination of highly efficient spiral sampling with dynamic compressed sensing results in major acceleration for 4D flow MR imaging, which allows comprehensive assessment of abdominal vessel hemodynamics in a single breath hold.

Dynamic Compressive Sensing of Time-Varying Signals Via Approximate Message Passing

In this work the dynamic compressive sensing (CS) problem of recovering sparse, correlated, time-varying signals from sub-Nyquist, non-adaptive, linear measurements is explored from a Bayesian perspective. While there has been a handful of previously proposed Bayesian dynamic CS algorithms in the literature, the ability to perform inference on high-dimensional problems in a computationally efficient manner remains elusive. In response, we propose a probabilistic dynamic CS signal model that captures both amplitude and support correlation structure, and describe an approximate message passing algorithm that performs soft signal estimation and support detection with a computational complexity that is linear in all problem dimensions. The algorithm, DCS-AMP, can perform either causal filtering or non-causal smoothing, and is capable of learning model parameters adaptively from the data through an expectation-maximization learning procedure. We provide numerical evidence that DCS-AMP performs within 3 dB of oracle bounds on synthetic data under a variety of operating conditions. We further describe the result of applying DCS-AMP to two real dynamic CS datasets, as well as a frequency estimation task, to bolster our claim that DCS-AMP is capable of offering state-of-the-art performance and speed on real-world high-dimensional problems.