基于改进自适应协同滤波算法的混沌信号去噪

Chaotic Signal Denoising Based on Improved Adaptive Collaborative Filtering Algorithm

摘要: 协同滤波去噪算法能充分利用混沌信号的自相似结构特征，具有良好的去噪性能。本文针对传统的自适应协同滤波算法中相似块提取不灵活以及聚合重构过于简单等问题，通过错位搜索的方法优化了相似块的提取，并采用动态时间归整算法(DTW)对聚合重构部分进行了改进。仿真实验的结果表明，在不同的噪声水平下，本文提出的方法均优于传统的自适应协同滤波算法。相较于小波滤波、高斯滤波以及经验模态分解等去噪方法，本文提出的方法在处理长期的混沌信号时具有更好的表现。

Abstract: The collaborative filter denoising algorithm can make full use of the self-similar structure characteristics of chaotic signals and has good performance. In this paper, for the problems of inflexible extraction of similar blocks and oversimplified aggregate reconstruction in the traditional adaptive collaborative filtering algorithm, the dislocation search method is used to optimize the extraction of similar blocks, and the Dynamic Time Warping (DTW) is used to improve the aggregation reconstruction. The simulation results show that the proposed method is superior to the traditional adaptive collaborative filtering algorithm under different noise levels. Compared with wavelet denoising, Gaussian filtering and empirical mode decomposition, the proposed method has better performance in dealing with long-term chaotic signals.