冲击噪声与持续性稳态噪声混合条件下的有源控制研究

The mixed noises which are combined by impulse and continuous steady state noise (called mixed noise for short) do exist widely in the industrial and military environments, the annoyance and harm of mixed noise is much severer than the noises with relatively single characteristics, therefore active control of mixed noise has important realistic demand, and is an inevitable issue for the engineering application of active noise control. Much attention is paid and great progress has been made on the active control of the continuous steady noise and the impulse noise. However, little work is presented to address the problem about active control of mixed noises. This project aims to find the appropriate methods and technologies for active control of mixed noises and focuses on three critical problems: (1) design the efficient adaptive algorithms for active control of mixed noise. The stability, convergence speed and the steady state performance of the designed algorithms should outperform the existing algorithms, the theoretical analysis of the step-size boundary, convergence speed and steady state error should also be derived; (2) study the suitable hybrid feedforward and feedback structures active control of mixed noise, which should alleviate the influence between the feedforward and feedback structures, guarantee the stability of the whole structure. The stability conditions and noise reduction performance of the hybrid structure should be analyzed theoretically; (3) validate the feasibility of active mixed noise control in free field and real mixed noise environments, obtain the results including noise reduction, bandwidth and system stability.

许多工业和军事噪声是冲击噪声和持续性稳态噪声兼而有之的“混合噪声”，混合噪声的总烦恼度与危害明显高于单一噪声，因此混合噪声的有源控制研究具有重要的现实需求，是有源噪声控制中无法回避的问题。本申请针对冲击噪声和持续性稳态噪声兼有的混合噪声的有源控制，重点研究三个关键问题：(1)设计有源控制的自适应算法，该算法在稳定性、收敛速度与降噪量指标上要优于现有算法直接用于混合噪声的结果。同时给出该算法的步长界限、收敛速度、收敛后误差等的理论分析。(2)研究适于混合噪声的有源控制前反馈组合结构，该结构能减轻前反馈结构间的相互影响，使得两者能够独立工作并保持稳定，理论上分析出该结构在混合噪声条件下的稳定性条件及可达到的降噪量。(3)检验混合噪声有源控制实际应用的可行性,在自由声场和实际混合噪声环境进行有源控制实验，获取降噪量、降噪带宽、系统稳定性的结果。

冲击噪声和持续性稳态噪声兼而有之的“混合噪声”广泛存在于工业、建筑业及军事领域。本项目针对混合噪声的有源控制在以下三个方面取得了研究进展：首先，在有源噪声控制算法方面，研究发现已有算法在控制混合噪声时会出现稳定性不足或收敛速度与降噪量不佳等问题，对此提出了一种改进的有源控制算法，该算法通过最小化后验误差对数变换的加权和，并考虑交换误差，在降噪量和收敛速度上都优于现有算法；为了使自适应反馈系统能明确调节噪声放大量，本项目研究了一种自适应算法，用实对称Toeplitz矩阵代替FxLMS算法的标量泄漏因子，矩阵中的元素可以根据噪声放大约束显式地计算，结果表明所提算法能够更为直接精确地对指定频段内的噪声放大量进行调节，从而在降噪量与噪声放大量之间取得折衷。其次，在有源噪声控制系统结构方面，研究了前反馈组合结构用于持续性稳态混合噪声的控制，发现现有组合结构中前反馈间存在耦合问题，给实际系统应用造成困难，对此提出一种改进结构可以解耦合，既发挥前反馈结构间的互补优势，又能抑制组合后两者的耦合效应。最后，在有源噪声控制系统的实验研究方面，开发了一套多通道有源控制系统，在降噪耳机、降噪头靠、通风管道中进行了有源噪声控制的实验验证。三年来，本项目的开展获得了混合噪声有源控制的新知识，研究了混合噪声有源控制的新技术、新方法。