Search

This paper proposes two novel schemes of wideband compressive spectrum sensing (CSS) via block orthogonal matching pursuit (BOMP) algorithm, for achieving high sensing accuracy in real time. These schemes aim to reliably recover the spectrum by adaptively adjusting the number of required measurements without inducing unnecessary sampling redundancy. To this end, the minimum number of required measurements for successful recovery is first derived in terms of its probabilistic lower bound. Then, a CSS scheme is proposed by tightening the derived lower bound, where the key is the design of a nonlinear exponential indicator through a general-purpose sampling-controlled algorithm (SCA). In particular, a sampling-controlled BOMP (SC-BOMP) is developed through a holistic integration of the existing BOMP and the proposed SCA. For fast implementation, a modified version of SC-BOMP is further developed by exploring the block orthogonality in the form of sub-coherence of measurement matrices, which allows more compressive sampling in terms of smaller lower bound of the number of measurements. Such a fast SC-BOMP scheme achieves a desired tradeoff between the complexity and the performance. Simulations demonstrate that the two SC-BOMP schemes outperform the other benchmark algorithms., Accepted by IEEE Transactions on Communications

Recently, Internet of Things (IoT) is becoming an important enabler to achieve smart health care for patients, where electrocardiograph (ECG) analyzing has been widely studied. In order to realize long-term detection of ECG under the limited hardware and network resources especially in the power/computing constrained IoT scenarios, compressive sensing (CS) has attracted tremendous research efforts thanks to its high efficiency in sample collection and information retrieval. Most existing researches focus on reconstructing the original ECG signal from the received compressive samples, which however not only leads to high energy consumption but also suffers from unnecessarily huge computational complexity for full-dimensional recovery. To solve such problems, this paper aims at a task-cognizant sparse signal processing technique for retrieving the necessary information only. Specifically, a correlation-assist compression with adaptive length (CCAL) algorithm is developed, where ECG signal is compressed individually within each pseudo-period to retain the information of each separate heartbeat. In doing so, the compression length is dynamically varying, which is determined by calculating the correlation coefficients between adjacent fragments. In addition, a preprocessing filter is adopted to offset the performance loss resulted from compression. With the compressed heartbeat acquired through CCAL, we extract autoregressive (AR) coefficients directly in the compressed domain. Afterwards, an SVM classifier is used to identify the conditions of the ECG signal. Finally, real data from MIT-BIH arrhythmia database is used for validating the proposed techniques that work efficiently in moderate compression ratios, while achieving desired classification accuracy as the non-compression benchmark at much reduced computational complexity.

Grant-free non-orthogonal multiple access (NOMA) is considered as one of the most important methodologies for the machine-type communications (MTC). In the field of MTC, compressed sensing based multi-user detection (CS-MUD) has been recognized as an excellent candidate for joint user activity and data detection, since many users sporadically transmit short-size data packets at low rates. This article focuses on the CS-MUD problem in the up-link grant-free NOMA scenario, where users are (in)-active randomly in each time slot yet with high temporal correlation. First, we investigate the CS framework to fully extract the underlying side information in the temporal correlation and propose a novel CS-MUD algorithm. Then, to mitigate the performance degradation due to the imperfect channel estimation in practice, the proposed algorithm is further extended by utilizing the perturbed CS, where the impact of channel estimation errors is modeled as certain perturbation in the measurement matrix. Different from most of the state-of-the-art CS-MUD algorithms, both proposed algorithms can apply even in the absence of prior knowledge on the number of active users. Simulation results indicate that the proposed algorithms achieve better performance than the existing CS-MUD methods. Their convergence and complexity issues are also discussed theoretically and numerically.

Compressive sampling matching pursuit (CoSaMP) is an efficient reconstruction algorithm for sparse signal. When the signal is block sparse, i.e., the non-zero elements are presented in clusters, some block sparse reconstruction algorithms have been proposed accordingly. In this paper, we present a new block algorithm based on CoSaMP, called block compressive sampling matching pursuit (BlockCoSaMP). Compared with CoSaMP algorithm, the proposed algorithm shows improved performance when sparse signal is presented in block form. Restricted isometry property (RIP) of measurement matrix is an effective tool for analyzing the performance of the CS algorithm, and Block restricted isometry property (Block RIP) is the extension of traditional RIP. Based on the Block RIP, we derive the sufficient condition to guarantee the convergence of Block-CoSaMP algorithm. In addition, the number of required iterations is obtained. Finally, simulation experiments show that with the increase of block length, the performance of Block-CoSaMP algorithm approaches to that of block subspace pursuit (Block-SP) algorithm. When the block length and sparsity are small, the performance of Block-CoSaMP algorithm is better than that of the CoSaMP, l2/l1 norm and block orthogonal matching pursuit (BOMP) algorithms. Especially, when compared with CoSaMP and l2/l1 norm algorithms, the proposed algorithm exhibits more obvious performance gain.

As the key communication module of vehicle to everything (V2X) system, the V2X device needs to carry out a series of performance tests. To verify the large-scale communication performance of V2X device under different parameter configurations in real traffic scenarios, we launched a field test for V2X device. In the test, the performance of three different V2X devices are tested under three typical traffic scenarios respectively. Both line of sight (LOS) and none line of sight (NLOS) scenarios are included. The quality of service (QoS) of two-way link communication performance with vehicular message transmitting in different scenarios is detected and analyzed. We hope that the test results can provide the basis for the formulation of standards, which will promote the progress of the V2X industry and the improvement of V2X standards.

Modern Automatic Speech Recognition (ASR) systems can achieve high performance in terms of recognition accuracy. However, a perfectly accurate transcript still can be challenging to read due to disfluency, filter words, and other errata common in spoken communication. Many downstream tasks and human readers rely on the output of the ASR system; therefore, errors introduced by the speaker and ASR system alike will be propagated to the next task in the pipeline. In this work, we propose an ASR post-processing model that aims to transform the incorrect and noisy ASR output into a readable text for humans and downstream tasks. We leverage the Metadata Extraction (MDE) corpus to construct a task-specific dataset for our study. Since the dataset is small, we propose a novel data augmentation method and use a two-stage training strategy to fine-tune the RoBERTa pre-trained model. On the constructed test set, our model outperforms a production two-step pipeline-based post-processing method by a large margin of 13.26 on readability-aware WER (RA-WER) and 17.53 on BLEU metrics. Human evaluation also demonstrates that our method can generate more human-readable transcripts than the baseline method., Accepted in 2021 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2021)

Sum frequency generation microscopy is a label-free optical imaging technique with intrinsic molecular vibrational contrast for surface studies. Recent developments of compressive sensing broad-band hyperspectral SFG microscopy have demonstrated the potential application for imaging monolayer at metal surfaces with micrometer spatial resolution. Here is presented the capability of chemical imaging of spatially patterned monolayers of 1-octadecanethiol (ODT) and 16-methoxy-1-hexadecanethiol (MeOHT) molecules assembled on a copper surface. The spatial distributions of the monolayer with vibrational-spectral contrast are well-demonstrated at different frequency regions through reconstruction of the hypercube using a 3-dimensional total variation regularization algorithm (3DTV). Spatial-chemical distributions of each component are also reconstructed directly from the compressive measurements by endmember unmixing (CEU) scheme. Compared to 3DTV algorithm, the reconstruction from CEU shows spatial distribution of each component on the surfaces, and demonstrates the ability to characterize the domains of mixed-molecules on surfaces.

Recently, compressive sensing based classification, which is called compressive classification, has drawn a lot of attention, since it works directly in the compressive domain with low complexity. However, existing literatures assume perfectly known measurement matrix during compressive classification, which is impossible in many practical situations. In this paper, we focus on studying the performance of classification based on the compressive measurements under the perturbation, where the perturbation models the uncertainty of the measurement matrix. The upper and the lower bounds on the probability of misclassification of the compressive classification are evaluated by utilizing the Kullback-Leibler and Chernoff distances. Our results indicate that the performance depends on the variance of the perturbation when the perturbation obeys Gaussian distribution with zero mean. Moreover, compared with the one without perturbation, the performance of the compressive classification can be improved when the perturbation in each hypothesis is different from each other. Motivated by these observations, we design an improved sparse representation classification (SRC) framework by incorporating the perturbation item into the SRC framework and propose several enhanced SRC schemes for performance improvement. The experiments on MNIST datasets validate that the proposed SRC schemes outperform the existing standard SRC scheme.

Heterogeneous network (HetNet) is considered as a prime way to solve the limits of system capacity and broadband service coverage in traditional network. However, the deployments of small cells with varied sizes make the network topology more complicated. Self-organizing network (SON) technology, aiming to reduce the operational costs, is a significant technology in HetNet. One of the common use cases is to improve handover performance. In this paper, a handover optimization algorithm based on enhanced mobility state estimation (EMSE) is proposed. Considering both user equipment (UE) speed and handover types, the optimization algorithm based on EMSE combines selective Time-to-Trigger (TTT) and dynamic handover margin (HM)-adjusting in SON. Furthermore, the algorithm performance is compared with two different reference cases. Simulation results show that total handover failure has an obvious decline with our self-optimizing algorithm. Therefore, handover performance gets improved and UEs have better mobility robustness in HetNet through our algorithm.

We are presenting an anomaly detection system with compressive single-pixel camera and statistic metrics to identify existence of anomalies against a complex background. We illustrate the utility of our approach via both simulations and experiments.

A low-cost 3-D hyperspectral projector was built employing a digital micromirror device for spectral and spatial modulated structured light and we demonstrate its utility in recovering multicolor inhomogeneous participating media based on compressive sensing theory.

We propose a silicon-compatible plasmonic electro-optic modulator employing a silicon racetrack ring resonator coupled to a bus waveguide. A silicon-plasmonic hybrid phase shifter clad with electro-optic polymer is introduced to achieve high-speed performance and low energy consumption. Simulations show that the proposed modulator can achieve an extinction ratio of more than 15 dB at 1550-nm wavelength under a 1.2-V bias voltage. The misalignment tolerance and fabrication feasibility of the modulator are also discussed.

A broad-band sum frequency generation microscope has been developed for the study of molecular monolayers on surfaces. Because sum frequency generation is a vibrational spectroscopy based on a second-order optical process, it is uniquely sensitive to detecting a molecule's vibrational fingerprints specifically at interfaces. In this microscope, a structured illumination beam generated by a spatial light modulator is used to irradiate the sample with a series of sparsifying pseudorandom patterns. The spectra associated with each pattern are then input into a reconstruction algorithm to compressively recover the full hyperspectral image cube. As a proof-of-principle, this system performed molecule-specific imaging of a microcontact-printed self-assembled monolayer of alkanethiolate on copper. This hyperspectral compressive imaging effectively recovered both spatial and spectral surface features with compression greater than 80%, meaning more than a 5-fold decrease in acquisition time compared to traditional methods.

We propose and numerically demonstrate an ultrafast real-time ordinary differential equation (ODE) computing unit in optical field based on a silicon microring resonator, operating in the critical coupling region as an optical temporal differentiator. As basic building blocks of a signal processing system, a subtractor and a splitter are included in the proposed structure. This scheme is featured with high speed, compact size and integration on a silicon-on-insulator (SOI) wafer. The size of this computing unit is only 35 μm × 45 μm. In this paper, the performance of the proposed structure is theoretically studied and analyzed by numerical simulations.

We present a gonio-plenoptic imaging system that realizes a single-shot shape measurement for specular surfaces. The system is comprised of a collimated illumination source and a plenoptic camera. Unlike a conventional plenoptic camera, our system captures the BRDF variation of the object surface in a single image in addition to the light field information from the scene, which allows us to recover very fine 3D structures of the surface. The shape of the surface is reconstructed based on the reflectance property of the material rather than the parallax between different views. Since only a single-shot is required to reconstruct the whole surface, our system is able to capture dynamic surface deformation in a video mode. We also describe a novel calibration technique that maps the light field viewing directions from the object space to subpixels on the sensor plane. The proposed system is evaluated using a concave mirror with known curvature, and is compared to a parabolic mirror scanning system as well as a multi-illumination photometric stereo approach based on simulations and experiments.

A mode-selective grating reflector based on a hybrid plasmonic waveguide, together with a two-to-one coupling structure, is proposed and analyzed. The grating structure is designed to reflect only the fundamental even mode existing in the hybrid plasmonic waveguide, and to pass through the fundamental odd mode at a 1550-nm wavelength with an extinction ratio of ~16.7 dB. The simulated transmission spectra verify our analysis and the expected features of the mode-selective hybrid plasmonic Bragg grating.

The use of combining pico cell and macro cell, also known as heterogeneous networks (HetNets), is considered as a promising way to solve the limit of system capacity and complement coverage in traditional network HetNets which have changed the infrastructure of traditional cellular networks also bring challenges correspondingly. Mobility management for pico cells is challenging due to different cell size and random deployment, and an enhanced mobility state estimation is desired. In this paper, we propose a self-optimization algorithm for mobility state estimation. The enhancement adapts the weight automatically based on handover performance to weaken the cell crossing counting for pico cells in different network topologies. The simulation results will show the improvements in mobility state estimation (MSE).