一种用于主动降噪耳机的权重滤波误差信号滤波-x最小均方算法*

针对传统FxLMS算法前馈自适应主动降噪耳机系统因果性条件不足时在宽带噪声环境中产生的高频噪声抬升问题,该文引入权重滤波误差信号FxLMS算法用于抑制高频噪声的抬升,但该算法带来了低频降噪量不足问题。因此,进一步提出将固定系数混合控制器与权重滤波误差信号FxLMS算法结合,在解决高频噪声抬升问题的同时,保证了良好的低频降噪量。基于DSP平台实现了提出的主动降噪耳机方案。实验证明,该方案针对宽带和单频等噪声都取得了较好的降噪效果。     

管道模态波入射下HQ管降噪机理研究

随着涡扇发动机风扇直径增大转速降低,低频噪声控制问题引发格外关注。HQ管作为一种抗性降噪结构,在低频段拥有良好的降噪效果。本文首先建立了带HQ管的管道风扇噪声数值模型,并对声传播特性进行了细致的仿真分析,获得了与前期实验数据对比一致性良好的预测结果。最后通过不同HQ管排布螺旋角下降噪效果和声模态分解结果的对比分析,对HQ管降噪机理进行了探讨,发现HQ管对模态波降噪的主要模式是模态的反射与散射,并且不同于将入射模态散射为截止模态的传统认知,将入射模态散射为不占优的可传播模态可以大幅削弱占优模态的幅值,同样具有较为明显的降噪效果。     

鲁棒性有源头枕系统的设计方法*

为了降低舱室内休息区域乘客双耳处的低频宽带噪声,采用最优维纳滤波器的设计准则,设计出一种基于局域空间有源降噪的多通道前馈有源头枕系统。实验对比了人工头在不同位置处和不同转动角度时两通道和四通道有源头枕系统在人耳处的降噪效果;进一步地,针对人头转动时控制效果变差的情况,提出了一种低复杂度的鲁棒性算法。仿真和实验表明,通道数量增加可使该系统在人耳附近获得更大的降噪量,使用鲁棒性算法的头枕系统在人头转动时有更好的控制性能。     

光度测量图像智能阈值降噪方法

为了提高光度测量精度,通过对光度测量过程中的入射散粒噪声、天光散粒噪声、暗电流散粒噪声、读出噪声的来源进行详细分析,有针对性地描述了降低各种噪声的方法,并针对暗电流散粒噪声与读出噪声及其他噪声残留提出了一种基于灰度直方图统计的智能阈值降噪方法;大幅消除了噪声对测量结果的影响,通过对恒星测光图像降噪前后光度计算结果的精度比对发现,采用这种方法将光度测量精度从原来的0.314 Mv提高到0.092 4 Mv,精度改善效果明显,并且解决了使用CCD测量光度结果不稳定的问题,具有较好的应用价值。     

水下流激孔腔噪声前馈有源控制实验研究

水流流经腔体时引起的自激振荡会形成流激噪声,该文通过仿真和实验研究了水下流激孔腔噪声的主动声场控制。应用有源前馈控制方法,对比了参考信号的选取以及控制器中增加的泄露因子对控制结果的影响,并在水循环管路中对流激孔腔噪声前馈控制效果进行了实验验证。研究结果表明,前馈有源噪声控制方法可有效抑制水下流激孔腔噪声,在误差点和辅助观测点都取得了良好的降噪效果,最大降噪量大于8 dB,并且使用腔内信号作为参考的控制效果优于上游信号作为参考。     

基于深度学习的低频宽带隔声器件设计*

在实际应用中，通常需要将多个声人工结构单元进行组合来实现低频宽带的隔声降噪。这种组合结构往往参数较多，传统的设计方法很难对其进行高效的自动化设计。本文在集总参数模型的基础上，提出了一种基于深度学习的低频宽带隔声器件设计方法，并基于该方法完成了由9个二阶亥姆霍兹共鸣器单元组合而成的低频宽带隔声装置的设计。仿真结果表明，该隔声装置在158 Hz~522 Hz范围内均具有良好的隔声效果，从而验证了所提出方法的有效性。与传统方法相比，本文所提出的设计方法不仅减少了对设计者专业知识和设计经验的依赖，而且具有更高的设计效率，更强的通用性，未来有望进一步推广至其他声人工结构的设计领域。     

充液管路系统流体声与结构声的复合有源控制

采用基于谐频自适应控制算法的有源消声与消振系统对充液管路系统突出的低频线谱噪声进行有源控制实验研究.建立了泵水循环管路实验系统,在管路中安装有源消声器对流体声进行控制,在管路出口障板上采用8×8通道有源消振系统控制结构声辐射。开展的低频线谱噪声与振动有源控制实验结果表明,在50~200 Hz频带内,通过结合有源噪声与振动控制可在多数频点取得10 dB以上的降噪效果。针对该实验系统,通过分别控制流体声和结构声分析了两者的贡献.实验结果验证了有源消声与消振系统具有较好的降噪性能,各频点处流体声与结构声占比情况不同,需要综合控制流体声与结构声才可以取得显著的降噪效果。      

军用运输机机舱有源消声实验系统的设计与实现

为验证有源消声技术在军用运输机机舱低频噪声消除方面的有效性,设计和实现了一套机舱有源消声实验系统。采用“激振器+舱壁板”方式实现了飞机螺旋桨对机舱诱导噪声的声源模拟,设计了基于前馈控制结构的自适应有源噪声控制系统,构建了基于FX-LMS算法的自适应消声控制器,采用监测麦克风组对舱内空间的消声效果进行监测。实验结果验证了自适应有源噪声控制技术在军用飞机舱室消声降噪领域的有效性,并表明初、次级声源间距对自适应有源消声系统的消声效果具有重要的影响。     

一种能量比调控的有源噪声控制算法*

在前馈有源噪声控制系统中,建模信号与控制信号相互影响,建模信号的引入会导致系统降噪性能变差。为了减小建模信号的影响,提出一种基于能量比调控的次级通道在线建模有源噪声控制算法。利用控制过程与建模过程的误差能量比构造步长调控函数,分别调节控制过程与建模过程的步长值,从而减弱两者的相互影响。在次级通道建模过程中,对建模步长值采取分段调控的方法,并通过建模步长值的变化来调节建模信号,从而提升系统降噪性能。仿真结果表明,对于低频噪声信号的有源噪声控制,相比已有算法,提出的算法能获得较快的建模收敛速度和较高的降噪量。     

有源抗噪声护耳器H2/H∞优化方法研究

采用固定系数滤波器和自适应滤波器相结合来实现一种有源抗噪声护耳器系统。固定系数滤波器除了降低宽带噪声,其最主要的作用在于减小次级通道的不确定性。固定系数滤波器和次级通道滤波器都采用H_2/H_∞优化方法设计得到。仿真和实验结果表明:固定系数滤波器提高了自适应算法的稳定性,护耳器系统在宽带和周期性噪声环境中都取得了比较好的降噪效果。     

Model independent control of lightly damped noise/vibration systems

Feedforward control is a popular strategy of active noise/vibration control. In well-damped noise/vibration systems, path transfer functions from actuators to sensors can be modeled by finite impulse response (FIR) filters with negligible errors. It is possible to implement noninvasive model independent feedforward control by a recently proposed method called orthogonal adaptation. In lightly damped noise/vibration systems, however, path transfer functions have infinite impulse responses (IIRs) that cause difficulties in design and implementation of broadband feedforward controllers. A major source of difficulties is model error if IIR path transfer functions are approximated by FIR filters. In general, active control performance deteriorates as model error increases. In this study, a new method is proposed to design and implement model independent feedforward controllers for broadband in lightly damped noise/vibration systems. It is shown analytically that the proposed method is able to drive the convergence of a noninvasive model independent feedforward controller to improve broadband control in lightly damped noise/vibration systems. The controller is optimized in the minimum H2 norm sense. Experiment results are presented to verify the analytical results.     

Simulation study on active noise control for a 4-T MRI scanner

The purpose of this work is to study computationally the possibility of the application of a hybrid active noise control technique for magnetic resonance imaging (MRI) acoustic noise reduction. A hybrid control system combined with both feedforward and feedback loops embedded is proposed for potential application on active MRI noise reduction. A set of computational simulation studies were performed. Sets of MRI acoustic noise emissions measured at the patient's left ear location were recorded and used in the simulation study. By comparing three different control systems, namely, the feedback, the feedforward and the hybrid control, our results revealed that the hybrid control system is the most effective. The hybrid control system achieved approximately a 20-dB reduction at the principal frequency component. We concluded that the proposed hybrid active control scheme could have a potential application for MRI scanner noise reduction.     

Novel active noise-reducing headset using earshell vibration control

Active noise-reducing (ANR) headsets are available commercially in applications varying from aviation communication to consumer audio. Current ANR systems use passive attenuation at high frequencies and loudspeaker-based active noise control at low frequencies to achieve broadband noise reduction. This paper presents a novel ANR headset in which the external noise transmitted to the user's ear via earshell vibration is reduced by controlling the vibration of the earshell using force actuators acting against an inertial mass or the earshell headband. Model-based theoretical analysis using velocity feedback control showed that current piezoelectric actuators provide sufficient force but require lower stiffness for improved low-frequency performance. Control simulations based on experimental data from a laboratory headset showed that good performance can potentially be achieved in practice by a robust feedback controller, while a single-frequency real-time control experiment verified that noise reduction can be achieved using earshell vibration control.     

H∞ Feedback control and Fx-LMS feedforward control for car engine vibration attenuation

The attenuation of the low frequency noise produced by the car engine vibration has been the major concern of the automotive community with the objective to increase the comfort of the driver and the passengers. The chassis subframe is the part responsible for transmission of the engine vibration to the car body through the engine mounts. At the beginning of this paper, we briefly describe the test rig we used and which has been experimentally investigated by using system identification. Then the experimental results of the feedforward control strategy and the achieved performance are briefly presented. These results are then compared to the feedback control method utilizing H_∞ strategy which allows us to design a robust controller requiring less sensors. Furthermore, a broadband noise attenuation could be achieved by using a robust feedback controller in the real time test.     

Zero/pole placement control of periodic duct noise

A zero/pole placement controller (ZPPC) is presented here for active control of periodic duct noise. It uses sound pressure feedback instead of modal feedback. The controller is able to place closed-loop zeros in addition to closed-loop poles. A new adaptation law is proposed to help the controller to track parameter drifting and maintain good performance. The controller can combine with any feedforward scheme to form a hybrid (feedforward/feedback) active noise controller. It has an advantage of on-line identification of the secondary path without persistent excitations.     

Synthesis of a robust broadband duct ANC system using convex programming approach

A robust active controller using spatially feedforward structure is proposed for broadband attenuation of noise in ducts. To meet the requirements of performance and robust stability in the presence of plant uncertainties, an H2 cost function and an H(infinity) constrain are employed in the synthesis of the controller. The design is then converted into a convex programming problem using Q-parametrization and frequency discretization. An optimal controller that satisfies the quadratic cost functions and linear inequality constraints can be found by sequential quadratic programming. The optimal controller was implemented via a digital signal processor (DSP) and verified by experiments. Experiment results showed that the system attained 16.5 dB maximal attenuation and 5.9 dB total attenuation in the frequency band 200-600 Hz.     

Design of a broadband active silencer using μ-synthesis

A robust spatially feedforward controller is developed for broadband attenuation of noise in ducts. To meet the requirements of robust performance and robust stability in the presence of plant uncertainties, a μ-synthesis procedure via D–K iteration is exploited to obtain the optimal controller. This approach considers uncertainties as modelling errors of the nominal plant in high frequency and is implemented using a floating point digital signal processor (DSP). Experimental investigation was undertaken on a finite-length duct to justify the proposed controller. The μ- controller is compared to other control algorithms such as the H₂ method, the H_∞ method and the filtered-U least mean square (FULMS) algorithm. Experimental results indicate that the proposed system has attained 25.8 dB maximal attenuation in the band 250–650 Hz.     

IMPLEMENTATION OF A BROADBAND DUCT ANC SYSTEM USING ADAPTIVE SPATIALLY FEEDFORWARD STRUCTURE

An adaptive spatially feedforward algorithm is proposed for broadband attenuation of noise in ducts. Acoustic feedback generally exists in this active noise control structure. Munjal and Eriksson (1988 Journal of Acoustical Society of America84, 1086-1093) derived an ideal controller for the spatially feedforward structure. The ideal controller can be partitioned into two parts. The first part represents a repetitive controller that can be implemented by an infinite impulse response (IIR) filter, whereas the second part represents the dynamics of transducer that can be implemented by a finite impulse response (FIR) filter. In the paper, the IIR filter is merged with the original plant. The FIR filter is adaptively updated by the least-mean-square (LMS) algorithm to accommodate perturbations and uncertainties in the system. The proposed algorithm is implemented via a floating point digital signal processor and compared with other commonly used algorithms such as the Filtered-X LMS algorithm, the feedback neutralization algorithm, and the Filtered-U LMS algorithm. Experimental results show that the system has attained 15·7 dB maximal attenuation in the frequency band 200-600 Hz.     

基于扰动源分类控制的二维陀螺平台过顶稳定方法

针对二维陀螺平台方位瞄准线控制在过顶位置时因驱动轴和敏感轴存在非线性约束导致的不稳定问题,分析了不同类型扰动源对方位瞄准线稳定的影响及其随俯仰角变化的规律,提出了基于扰动源分类控制的过顶稳定方法。该方法采用反馈和前馈双通道复合控制结构,在过顶位置时基于控制结构自身消除陀螺测量噪声放大导致的内生力矩扰动,通过增加前馈通道的滤波环节,抑制横滚扰动高频分量引起的力矩扰动,解决了过顶位置时方位驱动轴震荡的问题,同时通过反馈通道和前馈通道分别实现对方位扰动和横滚扰动低频分量的有效隔离。仿真结果表明,该方法能够大幅衰减陀螺测量噪声和横滚扰动高频分量引起的方位电机力矩扰动幅值,增强系统稳定性。最后通过某二维陀螺平台进行了实验,过顶位置时瞄准线方位经受振动条件下的稳定精度由82.4 μrad 减小为44.6 μrad,经受摇摆条件下的隔离度由?14.54 dB提升至?27.85 dB。实验结果验证了该方法能够有效提升过顶位置方位瞄准线的扰动隔离性能。