基于总体平均经验模态分解的主动噪声控制系统研究

为了提高宽窄带混合噪声的消噪效果，本文提出一种基于总体平均经验模态分解（Ensemble empirical mode decomposition，EEMD）的主动噪声控制（Active noise control，ANC）系统，利用实时EEMD算法逐段将混合噪声分解成若干个固有模态函数（Intrinsic mode functions，IMF）分量.因为这些IMF分量的频带各不相同，所以实现了混合噪声中宽带分量和窄带分量的有效分离，独立进行ANC处理后成功解决了处理混合噪声时带来的“火花”现象，而且避免了传统混合ANC（Hybrid ANC，HANC）系统中频率失调的影响. EEMD算法也是对混合噪声的平稳化处理过程，因此当混合噪声中出现非平稳变化时，本文提出的系统也能保持较好的系统稳定性.通过不同噪声环境下进行仿真分析，提出的ANC系统比HANC系统具有更好的系统稳定性和更小的稳态误差.

A New ANC System Based on Ensemble Empirical Mode Decomposition

In order to obtain a better de-noising performance of mixture noise containing both wideband components and narrowband components, a new active noise control (ANC) system based on ensemble empirical mode decomposition (EEMD) is proposed in this paper. Real-time EEMD algorithm is used to decompose the mixture noise into several intrinsic mode functions (IMF) which have a different frequency range each other, so this decomposition can separate wideband components and narrowband components from the mixture noise adaptively. Each component controlled independently can not only process mixture noise without “firework”, but also avoid the frequency mismatch occurring in conventional hybrid ANC (HANC) system. The EEMD algorithm can smooth the mixture noise to make sure the proposed system has better stability when non-stationary phenomenon happens in mixture noise. Compared with HANC system, the proposed ANC system has better system stability and smaller steady-state error in processing different noise.