基于ANC系统非线性次级通道的FXAP凸组合算法研究

发布时间：2018-03-28 07:03

本文选题：有源噪声控制系统　切入点：仿射投影算法　出处：《大连海事大学》2017年硕士论文

【摘要】：现代社会的高速发展使噪声污染的问题日益突出,因此,噪声控制是当前迫切需要进行的工作。传统的无源噪声控制技术对高频噪声具有较好的控制效果,但对抑制低频噪声作用不大,而人类生活环境多数处于低频范围,因此有源噪声控制(ActiveNoise Control,ANC)技术便应运而生了,并且自适应ANC系统因具有自动追踪初级噪声统计特性,控制器特性可时变的特点成为抑制低频噪声的研究热点。在自适应ANC系统中,线性模型虽然能简化运算和分析,但不能实际地评估系统的性能,同时次级通道中的信号转换器、功率放大器、传感器在把数字信号转换模拟信号或者机械信号的过程中往往会出现非线性转换。很小的非线性对算法的性能有着不可忽视的影响。另外,影响自适应ANC滤波器的性能因素来自其自适应控制算法和自身的结构两大主要方面。为此,本文做了以下几方面的工作:(1)首先建立了单通道前馈自适应ANC系统模型,并分析了该系统的核心—自适应滤波器和其对应的自适应滤波算法,其中对滤波-X最小均方算法(FXLMS算法)、滤波-X归一化最小均方算法(FXNLMS算法)以及滤波-X仿射投影算法(FXAP算法)的性能做了简要地介绍和仿真对比。(2)着重研究在收敛速度和失调取得较好的平衡的FXAP算法,并在次级通道处引入饱和非线性函数g(·),从能量守恒角度深入研究单通道前馈自适应ANC系统中非线性次级通道下FXAP算法在实数范围内的稳态性能和瞬态性能以及计算要求,仿真实验验证了理论分析。(3)为了提高该系统中滤波器的性能,基于凸组合算法和非线性次级通道下FXAP算法的特点,提出了单通道前馈自适应ANC系统中FXAP凸组合算法和非线性次级通道下FXAP凸组合算法,并分别对前者在实数范围内的稳态均方误差进行理论推导,对后者的计算要求、平稳状态和非平稳状态、高斯噪声和Bernoulli-Gaussian脉冲噪声环境下的性能做了大量地介绍和仿真对比分析。
[Abstract]:With the rapid development of modern society, the problem of noise pollution is becoming more and more serious. Therefore, noise control is an urgent task at present. The traditional passive noise control technology has better control effect on high frequency noise. But it has little effect on the suppression of low frequency noise, and most of human living environment is in the low frequency range, so active noise control ANC (active noise Control) technology has emerged, and the adaptive ANC system has the statistical characteristics of automatic tracking primary noise. In adaptive ANC systems, the linear model can simplify the calculation and analysis, but it can not actually evaluate the performance of the system, at the same time, the signal converter in the secondary channel. Power amplifiers, sensors, in the process of converting digital signals into analog signals or mechanical signals, there is often a non-linear conversion. The small nonlinearity has an important effect on the performance of the algorithm. The main factors affecting the performance of adaptive ANC filter come from its adaptive control algorithm and its own structure. Therefore, the following work is done in this paper: firstly, the single-channel feedforward adaptive ANC system model is established. The core of the system-adaptive filter and its corresponding adaptive filtering algorithm are analyzed. The performance of FXLMS algorithm, filter X normalized least mean square algorithm (FXNLMS) and filter X affine projection algorithm (FXAP) are briefly introduced and compared with simulation. Speed and misalignment achieve a better balance of the FXAP algorithm, The saturation nonlinear function g (I) is introduced at the secondary channel. From the point of view of energy conservation, the steady-state and transient performance of the FXAP algorithm in the nonlinear secondary channel in the single-channel feedforward adaptive ANC system is studied, as well as the calculation requirements. The simulation results show that in order to improve the performance of filter in this system, the characteristics of convex combination algorithm and FXAP algorithm based on nonlinear secondary channel are verified. The FXAP convex combination algorithm in single-channel feedforward adaptive ANC system and the FXAP convex combination algorithm in nonlinear secondary channel are proposed. The steady-state mean square error of the former in the real range is derived theoretically, and the calculation requirements of the latter are given. The performances of stationary state and non-stationary state Gao Si noise and Bernoulli-Gaussian impulse noise are introduced and simulated.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TB535;TN713

【相似文献】