The performance guarantees of generalized orthogonal matching pursuit（ gOMP） are considered in the framework of mutual coherence. The gOMP algorithmis an extension of the well-known OMP greed algorithmfor compressed...The performance guarantees of generalized orthogonal matching pursuit（ gOMP） are considered in the framework of mutual coherence. The gOMP algorithmis an extension of the well-known OMP greed algorithmfor compressed sensing. It identifies multiple N indices per iteration to reconstruct sparse signals.The gOMP with N≥2 can perfectly reconstruct any K-sparse signals frommeasurement y = Φx if K 〈1/N（1/μ-1） ＋1,where μ is coherence parameter of measurement matrix Φ. Furthermore,the performance of the gOMP in the case of y = Φx ＋ e with bounded noise ‖e‖2≤ε is analyzed and the sufficient condition ensuring identification of correct indices of sparse signals via the gOMP is derived,i. e.,K 〈1/N（1/μ-1）＋1-（2ε/Nμxmin） ,where x min denotes the minimummagnitude of the nonzero elements of x. Similarly,the sufficient condition in the case of G aussian noise is also given.展开更多
针对工业环境中随机冲击干扰下滚动轴承微弱故障特征提取难题,提出一种基于自适应短时维纳滤波(Adaptive Short Time Wiener Filtering,ASTWF)和改进正交匹配追踪(Orthogonal Matching Pursuit,OMP)的滚动轴承故障特征提取方法。该方法...针对工业环境中随机冲击干扰下滚动轴承微弱故障特征提取难题,提出一种基于自适应短时维纳滤波(Adaptive Short Time Wiener Filtering,ASTWF)和改进正交匹配追踪(Orthogonal Matching Pursuit,OMP)的滚动轴承故障特征提取方法。该方法首先采用包络峭度和随余比(Random Shocks and Margin Ratio,RMR)作为联合判据,界定窗长界限并自适应确定STWF最优窗长参数,进而将随机冲击干扰从测试信号中分离出来;然后,利用立方包络自相关谱估计信号中周期频率,构造周期原子库,降低匹配原子冗余度;最后,利用相似性理论优化匹配追踪迭代终止条件,并结合周期原子库,实现弱故障冲击特征快速、准确提取。根据仿真信号和通过变速箱下线检测所得工程数据,可验证所提出方法可有效识别随机冲击干扰下的滚动轴承微弱故障特征。对比最小熵形态反卷积(Minimum Entropy Morphological Deconvolution,MEMD)方法对于随机冲击干扰下滚动轴承微弱故障特征提取效果,发现所提出方法具有更好的故障特征提取能力;与经典OMP方法相比,所提出改进OMP方法信号重构速度提升66%。展开更多
As the amount of data produced by ground penetrating radar （GPR） for roots is large, the transmission and the storage of data consumes great resources. To alleviate this problem, we propose here a root imaging algor...As the amount of data produced by ground penetrating radar （GPR） for roots is large, the transmission and the storage of data consumes great resources. To alleviate this problem, we propose here a root imaging algorithm using chaotic particle swarm optimal （CPSO） compressed sensing based on GPR data according to the sparsity of root space. Radar data are decomposed, observed, measured and represented in sparse manner, so roots image can be reconstructed with limited data. Firstly, radar signal measurement and sparse representation are implemented, and the solution space is established by wavelet basis and Gauss random matrix; secondly, the matching function is considered as the fitness function, and the best fitness value is found by a PSO algorithm; then, a chaotic search was used to obtain the global optimal operator; finally, the root image is reconstructed by the optimal operators. A-scan data, B-scan data, and complex data from American GSSI GPR is used, respectively, in the experimental test. For B-scan data, the computation time was reduced 60 % and PSNR was improved 5.539 dB; for actual root data imaging, the reconstruction PSNR was 26.300 dB, and total computation time was only 67.210 s. The CPSO-OMP algorithm overcomes the problem of local optimum trapping and comprehensively enhances the precision during reconstruction.展开更多
基金Supported by the National Natural Science Foundation of China(60119944,61331021)the National Key Basic Research Program Founded by MOST(2010C B731902)+1 种基金the Program for Changjiang Scholars and Innovative Research Team in University(IRT1005)Beijing Higher Education Young Elite Teacher Project(YET P1159)
文摘The performance guarantees of generalized orthogonal matching pursuit（ gOMP） are considered in the framework of mutual coherence. The gOMP algorithmis an extension of the well-known OMP greed algorithmfor compressed sensing. It identifies multiple N indices per iteration to reconstruct sparse signals.The gOMP with N≥2 can perfectly reconstruct any K-sparse signals frommeasurement y = Φx if K 〈1/N（1/μ-1） ＋1,where μ is coherence parameter of measurement matrix Φ. Furthermore,the performance of the gOMP in the case of y = Φx ＋ e with bounded noise ‖e‖2≤ε is analyzed and the sufficient condition ensuring identification of correct indices of sparse signals via the gOMP is derived,i. e.,K 〈1/N（1/μ-1）＋1-（2ε/Nμxmin） ,where x min denotes the minimummagnitude of the nonzero elements of x. Similarly,the sufficient condition in the case of G aussian noise is also given.
文摘针对工业环境中随机冲击干扰下滚动轴承微弱故障特征提取难题,提出一种基于自适应短时维纳滤波(Adaptive Short Time Wiener Filtering,ASTWF)和改进正交匹配追踪(Orthogonal Matching Pursuit,OMP)的滚动轴承故障特征提取方法。该方法首先采用包络峭度和随余比(Random Shocks and Margin Ratio,RMR)作为联合判据,界定窗长界限并自适应确定STWF最优窗长参数,进而将随机冲击干扰从测试信号中分离出来;然后,利用立方包络自相关谱估计信号中周期频率,构造周期原子库,降低匹配原子冗余度;最后,利用相似性理论优化匹配追踪迭代终止条件,并结合周期原子库,实现弱故障冲击特征快速、准确提取。根据仿真信号和通过变速箱下线检测所得工程数据,可验证所提出方法可有效识别随机冲击干扰下的滚动轴承微弱故障特征。对比最小熵形态反卷积(Minimum Entropy Morphological Deconvolution,MEMD)方法对于随机冲击干扰下滚动轴承微弱故障特征提取效果,发现所提出方法具有更好的故障特征提取能力;与经典OMP方法相比,所提出改进OMP方法信号重构速度提升66%。
基金supported by the Fundamental Research Funds for the Central Universities(DL13BB21)the Natural Science Foundation of Heilongjiang Province(C2015054)+1 种基金Heilongjiang Province Technology Foundation for Selected Osverseas ChineseNatural Science Foundation of Heilongjiang Province(F2015036)
文摘As the amount of data produced by ground penetrating radar （GPR） for roots is large, the transmission and the storage of data consumes great resources. To alleviate this problem, we propose here a root imaging algorithm using chaotic particle swarm optimal （CPSO） compressed sensing based on GPR data according to the sparsity of root space. Radar data are decomposed, observed, measured and represented in sparse manner, so roots image can be reconstructed with limited data. Firstly, radar signal measurement and sparse representation are implemented, and the solution space is established by wavelet basis and Gauss random matrix; secondly, the matching function is considered as the fitness function, and the best fitness value is found by a PSO algorithm; then, a chaotic search was used to obtain the global optimal operator; finally, the root image is reconstructed by the optimal operators. A-scan data, B-scan data, and complex data from American GSSI GPR is used, respectively, in the experimental test. For B-scan data, the computation time was reduced 60 % and PSNR was improved 5.539 dB; for actual root data imaging, the reconstruction PSNR was 26.300 dB, and total computation time was only 67.210 s. The CPSO-OMP algorithm overcomes the problem of local optimum trapping and comprehensively enhances the precision during reconstruction.
