| 基于压缩感知理论的时变信号压缩技术 |
| |
| 引用本文： | 郭亮,高宏力,黄海凤,张筱辰.基于压缩感知理论的时变信号压缩技术[J].西南交通大学学报,2015,28(3):511-516. |
| |
| 作者姓名： | 郭亮 高宏力 黄海凤 张筱辰 |
| |
| 基金项目： | 国家自然科学基金资助项目(51275426) |
| |
| 摘 要： | 机械设备在线状态监测与故障诊断系统中,为实现采集数据的远程传输和实时处理,需对数据进行压缩和滤波处理.通过分析机械振动信号时变性的特点,以压缩感知理论为基础,构造了满足受限等距性质(restricted isometric property, RIP)的稀疏变换矩阵和压缩感知矩阵;提出了基于压缩感知的时变信号压缩算法,并利用Lasso算法对压缩信号进行稀疏重构,恢复原始信号.采用不同类型的时变仿真信号和实测信号进行实验,对比了提出算法与现有算法的压缩与去噪效果.实验结果表明,新算法有更好的压缩去噪效果,当压缩比为40%时,能量保持率达到了95%以上,能满足工程实际需求.
|
| 关 键 词： | 压缩感知 振动信号 数据压缩 信号去噪 |
| 收稿时间： | 2014-08-21 |
Time-Varying Signal Compression Technology Based on Compressed Sensing |
| |
| GUO Liang,GAO Hongli,HUANG Haifeng,ZHANG Xiaochen.Time-Varying Signal Compression Technology Based on Compressed Sensing[J].Journal of Southwest Jiaotong University,2015,28(3):511-516. |
| |
| Authors: | GUO Liang GAO Hongli HUANG Haifeng ZHANG Xiaochen |
| |
| Abstract: | Data compression and de-noise play important roles in the real-time storage and remote transmission of vibration signal for the on-line monitoring and fault diagnosis of mechanical equipments. By analyzing the time-varying characteristics of vibration signals, the sparse transfer matrix and compressing sensing matrix which satisfying the restricted isometric property (RIP) were constructed. Then, the signal compression and de-noise method was proposed based on compression sensing. The Lasso algorithm was used to reconstruct the compressed vector to restore raw signal. The proposed method was applied to analyzing the simulation and spindle vibration signal and was compared with the wavelet-based technique. The experimental results show that the proposed method is more effective in data compression and de-noise, and when the signal compression rate is 40%, the energy retention rate exceeds 95% to meet industrial requirements. |
| |
| Keywords: | |
|
| 点击此处可从《西南交通大学学报》浏览原始摘要信息 |
|
点击此处可从《西南交通大学学报》下载全文 |
|
