基于稀疏分解的轨道移频信号去噪算法研究

发布时间：2018-05-28 06:32

本文选题：轨道移频信号 + 去噪算法　；参考：《计算机测量与控制》2014年09期

【摘要】：随着我国铁路的高速发展,轨道移频信号的检测译码技术受到广泛关注;然而实际采集的轨道移频信号不可避免地会混入大量的背景噪声和干扰,因此译码前需要去噪以提高译码的准确性;提出一种基于稀疏分解的轨道移频信号去噪算法,利用移频信号特点构建过完备原子库,采用粗细二阶段匹配追踪算法实现移频信号的噪声去除;将文章算法应用到主流的ZPW-2000轨道移频信号中,结果表明,该算法具有比小波阈值、经验模式分解算法更好的去噪性能,能够有效地去除低信噪比移频信号的噪声,且去噪后译码信噪比可提高10 dB,另外,采用粗细二阶段原子搜索算法显著降低了匹配追踪的运算量,满足实时性要求。
[Abstract]:With the rapid development of railway in China, the detection and decoding technology of track frequency shift signal is paid more and more attention. However, the actual track frequency shift signal will inevitably be mixed with a large number of background noise and interference. In order to improve the accuracy of decoding, a new de-noising algorithm for track frequency shift signals based on sparse decomposition is proposed, which uses the characteristics of frequency-shift signals to construct an over-complete atomic library. The noise removal of the frequency shift signal is realized by using the coarse and fine two-stage matching tracking algorithm, and the algorithm is applied to the mainstream ZPW-2000 track frequency shift signal. The results show that the algorithm has better denoising performance than the wavelet threshold and empirical mode decomposition algorithm. It can effectively remove the noise of the low SNR frequency-shift signal, and the signal-to-noise ratio of decode can be increased by 10 dB after denoising. In addition, the coarse two-stage atomic search algorithm can significantly reduce the computational complexity of matching tracing and meet the real-time requirements.
【作者单位】：西南交通大学交通信息工程及控制重点实验室;
【基金】：中央高校基本科研业务费专项资金(SWJTU11CX041;SWJTU12CX099)资助四川省杰出青年培育基金(2011JQ0027) 中国铁路总公司科技研究开发计划课题(2013X012-A-1,2013X012-A2)
【分类号】：TN911.73

【参考文献】