Earticle

Distributed compressed sensing (DCS) states that we can recover the sparse signals from very few linear measurements. Various studies about DCS have been carried out recently. In many practical applications, there is no prior information except for standard sparsity on signals. The typical example is the sparse signals have block-sparse structures whose non-zero coefficients occurring in clusters, while the cluster pattern is usually unavailable as the prior information. To discuss this issue, a new algorithm, called backtracking-based adaptive orthogonal matching pursuit for block distributed compressed sensing (DCSBBAOMP), is proposed. In contrast to existing block methods which consider the single-channel signal reconstruction, the DCSBBAOMP resorts to the multi-channel signals reconstruction. Moreover, this algorithm is an iterative approach, which consists of forward selection and backward removal stages in each iteration. An advantage of this method is that perfect reconstruction performance can be achieved without prior information on the block-sparsity structure. Numerical experiments are provided to illustrate the desirable performance of the proposed method.

Compressed sensing-based matching pursuit algorithms can estimate the sparse channel of massive multiple input multiple-output systems with short pilot sequences. Although they have the advantages of low computational complexity and low pilot overhead, their accuracy remains insufficient. Simply multiplying the weight value and the estimated channel obtained in different iterations can only improve the accuracy of channel estimation under conditions of low signal-to-noise ratio (SNR), whereas it degrades accuracy under conditions of high SNR. To address this issue, an improved weighted matching pursuit algorithm is proposed, which obtains a suitable weight value u_op by training the channel data. The step of the weight value increasing with successive iterations is calculated according to the sparsity of the channel and u_op. Adjusting the weight value adaptively over the iterations can further improve the accuracy of estimation. The results of simulations conducted to evaluate the proposed algorithm show that it exhibits improved performance in terms of accuracy compared to previous methods under conditions of both high and low SNR.

본 논문에서는 멀티콥터의 효율적 영상 전송을 위해 필요한 압축센싱 기법을 제안한다. 제안된 구조는 압축센싱에 기반한 데이터 용량을 줄이는 것에 중점을 둔다. 우선 Spectrum sensing의 기본원리를 설명하고 AMP(Approximate Message Passing)와 CoSaMP(Compressive Sampling Matched Pursuit)을 수학적 분석과 모의실험 결과를 통해서 비교한다. 또한 두 알고리즘을 계산시 간과 복잡도 관점에서 평가하고 멀티콥터 동작에 적합한 알고리즘을 제안한다. 본 논문의 실험결과는 AMP 알고리즘이 CoSaMP 알고리즘보다 계산시간이 적고 이미지 에러 확률도 낮다는 것을 보여준다.

This paper proposed a novel compressed sensing (CS) technique for an efficient video transmission of multi-copter. The proposed scheme is focused on reduction of the amount of data based on CS technology. First, we describe basic principle of Spectrum sensing. And then we compare AMP(Approximate Message Passing) with CoSaMP(Compressive Sampling Matched Pursuit) through mathematical analysis and simulation results. They are evaluated in terms of calculation time and complexity, then the promising algorithm is suggestd for multicopter operation. The result of experiment in this paper shows that AMP algorithm is more efficient than CoSaMP algorithm when it comes to calculation time and image error probability.

본 논문에서는 3D 방송의 기본적인 원리를 설명하고 압축 센싱(Compressed Sensing) 기술을 적용하여 3D 방송의 데이터 용량을줄이는 방식을 제안한다. 샘플링 이론과 압축 센싱 기술의 차이점을 설명하고 개념과 동작원리를 설명한다. 최근 제안된 압축 센싱의 복원 알고리즘인 SS-CoSaMP(Single-Space Compressive Sampling Matched Pursuit) 와 CoSaMP(Compressive SamplingMatched Pursuit)를 소개하고 이를 이용하여 데이터를 압축 복원하여 정확도를 비교한다. 두 알고리즘의 다양한 이미지 복원을 수행하고 계산시간을 비교한다. 결론적으로 낮은 복잡도를 갖는 3D 방송에 적합한 알고리즘을 판단한다.

This paper describes the basic principles of 3D broadcast system and proposes new 3D broadcast technology that reducesthe amount of data by applying CS(Compressed Sensing). Differences between Sampling theory and the CS technologyconcept were described. Recently proposed CS algorithm AMP(Approximate Message Passing) and CoSaMP(CompressiveSampling Matched Pursuit) were described. This paper compared an accuracy between two algorithms and a calculation timethat image data compressed and restored by these algorithms. As result determines a low complexity algorithm for 3Dbroadcast system.

Compressed sensing (CS) provides a new solution for the problems of requiring large amount of measurements data and long data acquisition time in radar application, and both issues also exist in ground penetrating radar (GPR). Aiming at this problem, we adopt impulse radar with CS framework, and transform the GPR imaging into sparse constraint optimization problem performed on time-domain sub-sampling in this paper. Specifically, it focuses on the impulse GPR imaging method based on CS under double underground targets condition containing noise and abundant clutter. Furthermore, the performance of matching reconstruction algorithms under the different signal to noise ratios (SNR), measurement dimensions and sparseness values is also presented. The experimental results show that CS algorithms based on matching reconstruction can obviously reduce measurement data, improve the image quality and make a better anti-noise performance. When SNR of measurement data is 1dB, the probability of accurate imaging can still reach 95%. So we may reasonably conclude that the regularized orthogonal matching pursuit algorithm has a better performance than the other matching algorithms.

To make sure human face can be more accurately identified in various poses, an identification method based on the characteristics human face image have been proposed in the thesis. First, it is a design for the optimal sampling matrix to acquire compressed measurement data; then it adopts iterative method based on the close loops of l2 and l1 in normal form to get an estimate on human face images, since the method based on norm l2 can measure the relevancy of human face images in the space and time of the continuous time point, and the method based on norm l1 mainly uses the modified total variation method and basis pursuit noise-reduction method. The simulation design has suggested that the method adopted in the thesis can achieve a higher human face identification rate and a remarkable promotion effect.

The traditional image fusion algorithm completed the fusion based on all pixel information. The time and space requirements are higher. The improved fusion algorithm used the theory of compressed sensing (CS) for the processing of remote sensing image fusion. Firstly, the source images using wavelet transform for sparse representation, then, the improved fusion algorithm used the observation matrix for image dimension measurement, and completed the image fusion in CS domain. Finally, the algorithm used the improved OMP algorithm to reconstruct the fused image. The improved fusion algorithm is only applied with a few measurement data of the compressed sensing, and overcomed the shortcomings of traditional pixel level fusion, the fusion algorithm achieved good experimental data.

Image super resolution reconstruction has important significance in remote sensing image feature extraction and classification etc.. Because the remote sensing image size is larger, it is difficult to super resolution reconstruction using multiple images, the compressed sensing (CS) theory was introduced into the super-resolution reconstruction. Algorithm designed the low pass filter to reduce the sample correlation matrix and wavelet, at the same time, the algorithm selects the partial Hadamard-matrix as the measurement matrix, it has faster reconstruction speed and low storage requirements, which ensure that the image reconstruction keep with the RIP criterion of compressed sensing theory . Finally, this paper realizes the remote sensing image super resolution reconstruction through the improved iterative algorithm. Experiments show that the reconstructed images of the PSNR value has increased, the reconstructed image has a better visual effect.

A large number of pilots are utilized to acquire channel information in traditional channel estimation for Orthogonal Frequency Division Multiplexing (OFDM) system, which leads to lower spectrum efficiency. For exploiting the sparse channel characteristics of 3GPP multipath channels, we employ the Compressed Sensing (CS) approach for channel estimation. Two CS-based recovery algorithms, Orthogonal matching pursuit (OMP) algorithm and Compressive sampling matching pursuit (CoSaMP) algorithm, are considered in this paper. The Bit error rate (BER) and Mean squared error (MSE) performance using traditional least square(LS), and two CS-based algorithms are given. Simulation results demonstrate that the CoSaMP algorithm achieves best performance with fewer pilots among three algorithms under 3GPP channels.

The block compressed sensing has brought forth the problem that the reconstructed image is of lower quality compared with that of the compressed sensing. A new method is proposed in this paper, named as Block Compressed Sensing of Self-adaptive Measurement and Combinatorial Optimization, which capably solves the problem. According to different sparsity of each image block, we firstly measure the blocks by using different projections; then, we choose measurement with the optimal reconstruction as the final measurement. Eventually, reconstruct the original image using the optimal measurement we got. The proposed method outperforms the compressed sensing in terms of real-time and better reconstruction quality is achieved than the block compressed sensing. Our experimental results verify the superiority of the proposed method.

For distributed optical fiber pipeline pre-warning system, the sampling rate used is very high and thus huge data will be generated, which makes it difficult to transfer and store. Compressive sensing is a new compressed sampling method in the field of signal processing which compresses and samples the signal simultaneously. In this paper, an adaptive compressive sensing method is presented for compression and reconstruction of distributed optical fiber pipeline data. First, partial reconstruction based detection method is used to detect whether a hazardous event happened, then different compression ratios are taken for different classes of signal thereby increasing the compression ratio. In signal reconstruction phase, a sparsity determination algorithm is used to determine the sparsity of different segment of the signal, and then wavelet tree combined with CoSamp algorithm is adopted to reconstruct the signal. The adaptive compression algorithm improves the compression ratio and the sparsity determination in reconstruction phase can determine the sparsity of each segment when the signal varies without prior knowledge of the sparsity of the signal. Experimental results show that, the proposed algorithm can obtain higher reconstruction accuracy at a relatively high compression ratio. Furthermore, location simulation shows that the reconstructed signal by the proposed method is effective for danger signal positioning.

Compressed sensing is an innovative technology, which provides a new sampling mode. The key problem in compressed sensing is the construction of sensing matrix, which has an important influence on the signal sampling and reconstruction algorithm. At present, in most cases the sensing matrix is a random structure, and is difficult to realize due to its huge storage in practical applications. In this paper, we introduce a novel deterministic construction of sensing matrix via Q-matrix, which is calculated by solving the N-queens problem. The proposed sensing matrix has good orthogonality and circularity. Using the circularity of Q-matrix, we can construct sensing matrix for compressed sensing. A large number of simulation results show that the proposed sensing matrix in this paper can obtain a better quality of the reconstructed image, and it is easily realized owing to its cyclic characteristic.

Random equivalent sampling (RES) can composite a waveform with high equivalent sampling rate from multiple low speed sampling sequences. In practical application, the performance of RES signal reconstruction would be degraded by the non-uniform distribution of sampling time. Compressed sensing (CS) theory is adopted to reconstruct RES samples, which could mitigate the inherent coherence of sampling time. However, the CS reconstruction algorithm is sensitive to the signal sparsity level that is unknown in the reconstruction stage. In this paper, we propose a redundancy reduction algorithm for CS base RES signal reconstruction that can ensure reconstruction accuracy while reducing the number of random samples. The experimental results are reported to evaluate the performance of the proposed algorithm.

This paper proposes a new method of efficient image reconstruction based on the Modified Frame Reconstruction Iterative Thresholding Algorithm (MFR ITA) developed under the compressed sensing (CS) domain by using total variation algorithm. The new framework is consisted of three phases. Firstly, the input images are processed by the multilook processing with their sparse coefficients using the Discrete Wavelet Transform (DWT) method. Secondly, the measurements are obtained from sparse coefficient by using the proposed fusion method to achieve the balance resolution of the pixels. Finally, the fast CS method based on the MFR ITA is proposed to reconstruct the high resolution image. The proposed method achieved superior results on real images, and demonstrate qualitative improvements in terms of PSNR and SSIM values. Furthermore, achieved good reconstruction SNR in the presence of noise.

The measurement matrix of compressed sensing has a significant impact for sampling and reconstruction algorithm of the original signal. At present, the majority of the measurement matrix is randomly constructed, and it is difficulty for hardware implementation in the practical applications. In this paper, we use the sparse characteristic of parity- check matrix of LDPC codes, construct measurement matrix based on QC-LDPC (Quasi-cyclic low-density parity-check) matrix, which is a structural and sparse deterministic measurement matrix. The simulation results show that, the measurement matrix is proposed in this paper not only can obtain a better reconstructed image quality, but also it can reduce the complexity of hardware implementation for quasi-cyclic.

Wireless Sensor Networks (WSNs) have characteristics of large size, limited resources, large amount of transmission data, and so on. In order to reduce the redundancy of sensed data and decrease network data traffic. We applied CS to clustered structure, proposed Low-Latency Compressed Sensing model (LLCS) which is based on the spatial-temporal correlation of sensed data, the model is also capable of processing sparse abnormal events which is a crucial feature in WSNs. We analyzed the relationship between compression ratio and sampling rounds and verified the abnormal event processing method. The results of simulation experiments using the real data show that LLCS could reduce data transfer volume significantly and process abnormal readings effectively.

A multi-target localization algorithm based on compressed sensing was proposed in this paper. The issue of multi-target localization was transformed into compressed sensing. The algorithm greatly reduced the amount of wireless network’s communication data by transferring most of the computing work to the central server. This method made full uses of the priori information of the signal and the support set. It combined Kalman filter with Bayesian compressed sensing to improve the localization accuracy and noise immunity. Simulation results showed that the proposed method has good noise immunity, robustness and localization accuracy compared with traditional localization methods.