Multichannel Active Noise Control Research Articles

Enhanced selective delayless subband algorithm independent of primary disturbance configuration for multi-channel active noise control system in vehicles

The selective delayless subband structure stands out as a promising algorithmic choice for the multi-channel active control of vehicle interior noise, particularly in the context of road noise. This type of algorithm reduces the eigenvalue spread of the autocorrelation matrix of the signal by decomposing the signal into subbands, and the desired subbands are activated selectively, thus achieving a significant performance improvement while consuming less computational resources compared with the traditional fullband algorithms. Nevertheless, the effectiveness of the subband algorithm appears to be closely tied to the configuration of the primary disturbance signal. In cases where the primary disturbance signal encompasses both broadband and tonal components, the performance advantage of the subband algorithm becomes constrained. In this paper, we conduct a thorough comparative analysis of two extensively employed subband algorithms, namely the Morgan and the Milani methods, through simulations employing data obtained from a multi-channel active noise control (ANC) headrest system. The results indicate that the Morgan method outperforms the Milani method when configured optimally. Subsequently, we propose an enhanced version of the subband algorithm based on the Morgan method. The enhanced algorithm incorporates an additional sinusoidal noise canceller (SNC) subsystem and a narrowband active noise control (NANC) subsystem based on local secondary path modeling to address tonal components, while the selective subband structure is employed for controlling broadband components. In addition, the computational complexity of the algorithm is analyzed. The effectiveness of the proposed algorithm is validated through numerous simulations and the ANC tests conducted on an actual vehicle utilizing the multi-channel ANC headrest system. The performance of the proposed algorithm surpasses that of traditional selective subband and fullband algorithms, and balanced binaural noise reduction is achieved.

What is behind the meta-learning initialization of adaptive filter? — A naive method for accelerating convergence of adaptive multichannel active noise control

Active noise control (ANC) is a typical signal-processing technique that has recently been utilized extensively to combat the urban noise problem. Although numerous advanced adaptive algorithms have been devised to enhance noise reduction performance, few of them have been implemented in actual ANC products due to their high computational complexity and slow convergence. With the rapid development of deep learning technology, Meta-learning-based initialization appears to become an efficient and cost-effective method for accelerating the convergence of adaptive algorithms. However, few dedicated Meta-learning algorithms exist for adaptive signal processing applications, particularly multichannel active noise control (MCANC). Hence, we proposed a modified Model-Agnostic Meta-Learning (MAML) initialization for the MCANC system.11The code of the proposed method is available on https://github.com/ShiDongyuan/Meta_Learning_for_MCANC. Additional theatrical research reveals that the nature of MAML, when applied to signal processing, is the expectation of a weight-sum gradient. Based on this discovery, we devised the Monte-Carlo Gradient Meta-learning (MCGM) algorithm, which employed a more straightforward procedure to accomplish the same performance as the Modified MAML algorithm. Furthermore, the numerical simulation of ANC using raw noise samples on measured paths validates the efficacy of the proposed methods in accelerating the convergence of the multichannel-filtered reference least mean square algorithm (McFxLMS).

The Effectiveness of Least Mean Squared-Based Adaptive Algorithms for Active Noise Control System in a Small Confined Space

Active noise control (ANC) is a technique applied to eliminate an unwanted sound by superposing a signal of equal amplitude and opposite phase, sometimes defined as an anti-noise signal, computed through an adaptive algorithm. The study described herein aims to evaluate and compare the performance of some of the most popular algorithms based on the least mean squares (LMS) approach applied to a multichannel active noise control system in a small, enclosed space. The comparison is conducted through an experimental evaluation of the ANC algorithms’ performance, carried out on a tractor cabin in a hemi-anechoic chamber, generating the unwanted sound field using a dodecahedron sound source placed outside the enclosure, emitting narrowband and broadband signals. The experimental analysis and the comparison with the results obtained in a free field condition have made it possible to show certain practical limitations when implementing the algorithms. The results show that the feed-forward systems allow for greater stability, avoiding the acoustic feedback from the control loudspeakers to the reference microphone when this is outside the cabin, while the feedback system is the slowest configuration to converge, requiring an internal modeling of the reference signal. With random signals, the feed-forward systems concentrate their performance in the range above 500 Hz, while the feedback system becomes ineffective.

A combined method for multi-channel active noise control based on online adaptive estimation and offline convolutional neural network

Multi-channel active noise control (ANC) has been widely used to attenuate low-frequency noise in relatively large spatial areas. Traditional multi-channel systems are achieved by optimizing the controller weights with adaptive algorithms that minimize the power of all error signals. However, nonlinearity in ANC systems can significantly degrade the performance of conventional adaptive algorithms, which is even more severe in multi-channel systems. Researchers suggested using neural networks (NN) to deal with the nonlinear problems. However, applying NN to multi-channelANC systems is challenging in real-time processing because of the high computational burden. Therefore, a multi-channel ANC method that combines NN and adaptive method is proposed. The proposed controller consists of both linear and nonlinear parts, where the linear part is estimated adaptively to track the change of primary noise in real applications while the nonlinear part is modeled offline with a small-scale convolutional NN (CNN). This method has the advantages of solving the nonlinear problem in ANC systems and achieving the capability in tracking primary noise source positions. At last, the performance of the proposed algorithm is verified through simulation experiments.

A hat-shape error microphone array for user-friendly spatial active noise control

Spatial active noise control (ANC) has been developed in the past decades to reduce unwanted acoustic noise over a desired region by generating an anti-noise field with secondary loudspeakers. Unlike conventional multi-channel ANC systems, the integration of spherical harmonic-based sound field analysis allows the ANC system to reduce noise over a continuous region surrounding the user's head rather than discrete points. However, the use of a spherical error microphone array on the boundary of the region for spherical harmonic analysis poses a limitation as it obstructs user access to the controlled area. In this work, we introduce a hat-shaped error microphone array design which can be placed above the listener's head. The proposed error microphone array can achieve spherical harmonic-based residual noise field analysis and is more user-friendly, allowing unrestricted user head movement. We construct the hat-shaped error microphone array and present the results of ANC experiments conducted using the array.

A low-complexity multi-channel active noise control system using local secondary path estimation and clustered control strategy for vehicle interior engine noise

Conventional multi-channel active noise control (ANC) system based on the adaptive notch filtered-x least mean square algorithm is typically used for active control of vehicle interior engine noise with prominent harmonic characteristics. However, due to the large estimated secondary path length and the centralized control strategy, the system has high computational complexity and fragile stability. To alleviate this problem, an efficient multi-channel ANC system using the local secondary path (LSP) estimation and clustered control strategy is proposed in this paper. Based on the modified LSP estimation method, the system may utilize a set of low-order but accurate LSP models to simplify the convolution operations of reference filtering. Besides, the proposed clustered control strategy combines the advantages of centralized and decentralized control strategies, and thus the system not only enjoys good noise attenuation performance and stability, but also its computational burden is further greatly reduced. Numerical simulations are performed to evaluate the noise attenuation performance and frequency mismatch effect of the proposed multi-channel ANC system. In addition, a series of real vehicle ANC experiments are conducted. The results show that, under stationary work conditions, the interior overall, 2nd order, 4th order, 6th order engine noises may be attenuated by 12.91 dB(A), 32.98 dB(A), 15.34 dB(A) and 8.74 dB(A), respectively, and under nonstationary work conditions, the maximum overall noise reductions at the four error microphones may reach 6.40 dB(A), 6.14 dB(A), 12.48 dB(A) and 11.07 dB(A), respectively. This fully confirms the effectiveness of the proposed multi-channel ANC system in real-world applications.

Reference Microphone Free Active Headrest System for Noise Containing Varying Frequency Tonal Components

In this brief, the design and implementation of an active headrest system are given. An efficient narrowband multichannel active noise control (ANC) algorithm known as filtered-e weighted-frequency Fourier linear combiner least mean square (FE-WFLC-LMS) is implemented in the proposed active headrest system to effectively control the undesired acoustic noise without hampering the desired sound such as speech and alarms. The computational complexity analysis along with simulation and real-time experiments are provided to demonstrate the efficiency of the proposed algorithm. Through simulation study, it is shown that the proposed multichannel FE-WFLC-LMS algorithm is capable of controlling the noise containing variable tonal frequencies effectively without use of reference microphone. In real-time experiments, it is observed that the proposed algorithm is able to track the frequency change to effectively control the noise without using any reference microphone.

A new variable spatial regularized FxLMS algorithm for diffusion active noise control

Distributed multi-channel active noise control (ANC) systems attract a lot of attention due to the reduced computational complexity than centralized control methods and improved stability than decentralized control methods. However, the combination of controllers within a neighborhood in a diffusion manner introduces an estimation bias and may degrade the control accuracy. This is because the secondary sources and error microphones of an ANC system are usually physically placed at different locations and the optimal solution to each controller is different. In this paper, a new diffusion filtered-x least mean squares algorithm (Diff-FxLMS) has been developed that balances combination strength and estimation bias via a variable spatial regularization. The mean squares error criterion subject to a bias constraint is used such that the spatial regularization parameter could be adapted according to the penalized Lagrangian. A detailed performance analysis is carried out, based on which user parameters can be selected automatically. Performance of the proposed variable spatial regularized Diff-FxLMS (VSR-Diff-FxLMS) algorithm and theoretical analysis is verified by simulations.

A Robust Adaptive Filter for Diffusion Strategy-based Distributed Active Noise Control

Active noise reduction in the existential of impulsive noise poses to be a serious challenge in a distributed scenario. The conventional centralized multichannel active noise control (MANC) and diffusion filtered-x least mean square (F-xLMS) based distributed active noise control (DANC) systems exhibit deteriorated performance in the existence of impulse noise and outliers. In this regard, a diffusion least logarithmic hyperbolic cosine function (D-Llncosh) based algorithm for DANC system is developed with the goal of mitigating noise in such an environment. It utilizes natural logarithmic of hyperbolic cosine function integrating the mean square error (MSE) and mean absolute error (MAE) as the cost function depending upon the parameter value. The weight update rule of the introduced D-Llncosh-based DANC system is hereby derived. The convergence analysis of the suggested method is also carried out. The proposed system is validated for various types of impulsive noise under different scenarios. The extensive simulation results assert the efficacy of the posited DANC system in comparison with the existing techniques. The suggested method achieves quicker convergence and approximately a 1–5 dB improvement is the noise reduction for different noise scenarios as evident from the obtained results.

Computation-efficient solution for fully-connected active noise control window: Analysis and implementation of multichannel adjoint least mean square algorithm

The multichannel active noise control (MCANC) system, in which multiple reference sensors and actuators are used to enlarge the noise-cancellation zone, is widely utilized in complex acoustic environments. However, as the number of channels increases, the practicality decreases due to the exponential rise in computational complexity. This paper, therefore, revisits the adjoint least mean square (ALMS) algorithm and its multichannel applications. The computational analysis reveals that the multichannel adjoint least mean square (McALMS) algorithm11The code is shared on https://www.mathworks.com/matlabcentral/fileexchange/127554-multichannel-adjoint-lms-algorithm-for-multichannel-anc. has a significantly lower computation cost when implementing the fully connected active noise control (ANC) structure. In addition to this advantage, the theoretical analysis presented in this paper demonstrates that the McALMS algorithm can achieve the same optimal solution as the standard adaptive algorithm without the assumptions of input independence and white Gaussian noise. In addition, a practical step-size estimation strategy based on the Golden-section search (GSS) method is proposed to predict the fast step size of the McALMS algorithm. The numerical simulations in a multichannel ANC system demonstrate the effectiveness of the McALMS algorithm and validate the derived theoretical analysis. Furthermore, the McALMS algorithm with proposed step-size approach is used to implement a multichannel noise cancellation window that achieves satisfactory global noise reduction performance for tonal, broadband, and even real-world noises.

An adaptive constrained multi-channel active noise control filter design approach using convex cone optimization

In current practical active noise control (ANC) applications, solutions to three main technical challenges need to be sought: the requirement of a stable controller when acoustic feedback paths exist, a larger quiet zone, and adaptation to a time-varying environment. The imposed controller stability constraints can significantly increase the computational requirements and negatively impact the adaptation of ANC filters in real time. Multi-channel systems are usually needed if a larger quiet zone is required while the coupling between different channels can greatly lower the convergence rate of adaptive algorithms. In the current work, a multi-channel adaptive constrained ANC filter design method is proposed. The recently proposed convex cone formulation for ANC design is adopted to reduce the computational time of solving the constrained ANC filter design problem. The optimal filter coefficients are then redesigned continuously using convex optimization when the environment is time-varying. The experimental result shows that, compared with the traditional leaky FxLMS method, the proposed method has a faster convergence rate and a better steady-state noise control performance.

Spatially selective active noise control systems.

Active noise control (ANC) systems are commonly designed to achieve maximal sound reduction regardless of the incident direction of the sound. When desired sound is present, the state-of-the-art methods add a separate system to reconstruct it. This can result in distortion and latency. In this work, we propose a multi-channel ANC system that only reduces sound from undesired directions, and the system truly preserves the desired sound instead of reproducing it. The proposed algorithm imposes a spatial constraint on the hybrid ANC cost function to achieve spatial selectivity. Based on a six-channel microphone array on a pair of augmented eyeglasses, results show that the system minimized only noise coming from undesired directions. The control performance could be maintained even when the array was heavily perturbed. The proposed algorithm was also compared with the existing methods in the literature. Not only did the proposed system provide better noise reduction, but it also required much less effort. The binaural localization cues did not need to be reconstructed since the system preserved the physical sound wave from the desired source.

Deep learning-based active noise control on construction sites

Although construction noise pollution has always been a severe issue for governments in metropolitan cities, there is no effective and practical solution to control it. Due to the high nonlinear and transient nature of machinery noises on construction sites, traditional active noise control (ANC) algorithms show a marginal ability to mitigate them. In this paper, a novel high-performance deep-learning-based feedforward ANC controller is proposed to attenuate construction-related noise by considering the delay and nonlinear behavior of acoustic devices. The developed network, with around 128,500 parameters, can be expanded to a multi-channel ANC method without increasing computational costs, and it is suitable for ANC in open space environments like construction sites. Broadband noise attenuation of around 8.3 dB was achieved of a wide variety of construction noises with minor degradation at very high-frequency ranges (7.5–8 kHz). The presented network outperformed traditional and state-of-the-art ANC algorithms.

Mitigating wind noise in multichannel active noise control systems

Multichannel active noise control (ANC) systems have been investigated to reduce outdoor low-frequency noise, such as power transformer noise and traffic noise. However, in outdoor environments, wind noise is inevitable and significantly degrades the noise reduction performance of active control systems. Despite broad adoption in windy environments, porous microphone windscreens cannot eliminate low-frequency wind noise in microphones, especially when the wind speed is high. Assuming wind noise is additive and uncorrelated with the primary noise to be controlled, this paper analyses how the wind noise detected by the reference microphone impacts the control filters and the residual noise of a multichannel ANC system based on the multichannel filtered-reference least mean squares algorithm (MFxLMS). By extending our recent research on wind noise suppression in single-channel active noise control systems [Y. Chu, et al., J. Acoust. Soc. Am. 152(6), 3340–3345 (2022)], a modified MFxLMS algorithm based on a new cost function is proposed. Simulations with real-recorded wind noise are performed to verify the theoretical analysis.

Design of causal preconditioning filters for adaptive filtering algorithms in real-time multi-channel active noise control applications

Least-mean-square (LMS) algorithm has been widely used in the area of active noise control (ANC). One practical concern of LMS-based algorithms is its slow convergence rate in multi-channel broadband noise control applications. To improve the convergence speed of traditional LMS algorithm, preconditioning filters were added to the LMS system in previous studies to remove the correlation between the reference signals and decouple the plant responses. However, the preconditioning filters implemented previously are usually obtained through singular value decomposition of the cross spectral matrix of reference signals and the plant response matrix, which does not lead to causal preconditioning filters and, thus, it can only be applied in spatial audio applications when delay is not an important concern, but cannot be implemented in real-time active noise control applications. In the current work, a method is proposed to obtain a casual multi-channel preconditioning filter for real-time active noise control applications through a numerically robust algorithm to perform a spectral factor decomposition to the reference signal cross spectral matrix and a minimum-phase and all-pass decomposition to the multi-channel secondary path. Simulation results show that, by applying the proposed preconditioning filter, the convergence speed of the LMS algorithm can be significantly improved.

Experimental tests of a multichannel active noise control system with single and multiple reference signals applied to the cabin of a tractor

A feedforward active noise control system represents a valid solution whenever it is necessary to attenuate a disturbing noise at low frequency. One example of application is the case of tractors, in which the agricultural operators are exposed for several hours per day to loud noise at low frequency, mainly generated by the engine. Generally, the feedforward systems make use of one reference sensor to measure the noise produced by the main source. However, this way of proceeding could be inadequate to describe the disturbing noise, especially when the noise is divided both in structure-borne and air-borne components. In this article, a comparison between the use of a single and multiple reference signals, processed in a multichannel active noise control system applied to the cabin of a motionless tractor, is proposed. In particular, an accelerometer positioned on the turbocharger is used to sense the vibration generated by the engine, while a microphone placed near the engine is used to measure the air-borne noise. The active system implements the filtered-X least mean squares (FXLMS) algorithm to produce the anti-noise signals. The results show that the use of two simultaneously acquired reference signals improves the performance of the active system.

Multichannel two-gradient direction filtered reference least mean square algorithm for output-constrained multichannel active noise control

Multichannel active noise control is a proven and long-standing technique for achieving a large zone of quiet in enclosures and open spaces. In practice, the multichannel active noise control system encounters more significant output saturation nonlinearity than the single-channel system as a result of the increased number of multiple output amplifiers, which not only degrades noise reduction performance but also compromises system stability. Conventional output-saturation solutions are mainly designed for single-channel applications, and their high computational complexity prevents them from being extended to multichannel systems. Hence, this paper proposes a novel two-gradient direction multichannel filtered reference least mean square (FxLMS) method that effectively avoids the saturation problem by restricting the output to the desired range and marginally increasing the computation over the conventional Multichannel FxLMS algorithm. In addition, this paper investigates the step sizes of the algorithm and proposes a computation-efficient variable step-size strategy to reduce further the steady-state error caused by the varying gradient directions. Finally, the efficacy of the proposed algorithm is demonstrated through numerical simulations using measured primary and secondary paths.

Distributed Active Noise Control Based on an Augmented Diffusion FxLMS Algorithm

Multichannel active noise control (ANC) systems have been widely investigated for low-frequency noise attenuation over a spatial region. Using a conventional centralized control strategy based on the multichannel filtered- <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</i> least mean square (F <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</i> LMS) algorithm has been demonstrated to be effective for multichannel ANC systems, but the high computational burden restricts its practical applications. Meanwhile, a decentralized control strategy suffers from stability problems although it has been successful in reducing the computational load. Recently, distributed control strategies, such as the multitask diffusion adaptation scheme, have been introduced to ANC systems and shown to mitigate the stability problems in decentralized systems. However, distributed ANC systems using the diffusion F <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</i> LMS algorithm require strong symmetry of acoustic paths because of the dependence on node-based adaptation and neighborhood-based combination. To overcome this limitation, this paper proposes an Augmented Diffusion F <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</i> LMS algorithm with neighborhood-based adaptation and node-based combination. A theoretical formulation and convergence analysis are presented and simulations are performed to compare the proposed algorithm with existing ones under different system configurations for tonal, multi-tonal, narrowband and broadband signals. Simulation results demonstrate that the proposed algorithm has the same noise reduction performance as centralized method even if the acoustic paths are strongly asymmetrical, which is superior over existing distributed multitask diffusion strategy.

Remote Microphone Technique for Active Noise Control Over Distributed Networks

Multichannel Active Noise Control (ANC) headrest systems have usually been designed with the objective of creating quiet areas at the passenger positions within the cabin of a public transport. Due to the high computational demands of dealing with multiple loudspeakers and multiple microphones by using a single centralized processor, and the convenience of a scalable system, a distributed implementation may be advisable. On the other hand, the addition of the remote microphone (RM) technique to the ANC headrest systems has recently allowed the creation of local zones of quiet at unachievable placements for the error sensors. This technique allows the quiet zone to be shifted to a desired location away from the physical sensor. However, to our knowledge, the implementation of an ANC headrest system with virtual microphones over a distributed acoustic sensor network has not been addressed. In this work, a distributed version of the multiple-error FxLMS (MEFxLMS) algorithm considering the RM method (RM-DMEFxLMS) is proposed by using an alternative formulation of the RM method. Through a set of simulations, it is found that on a non-constrained communication network, the proposed algorithm presents the same performance of its centralized version and provides a scalable and more versatile ANC system.

Deep MCANC: A deep learning approach to multi-channel active noise control

Traditional multi-channel active noise control (MCANC) is based on adaptive filtering and usually uses a separate control unit for each channel. This paper introduces a deep learning based approach for multi-channel active noise control (ANC). The proposed approach, called deep MCANC, encodes optimal control parameters corresponding to different noises and environments, and jointly computes the multiple canceling signals to cancel or attenuate the primary noises captured at error microphones. A convolutional recurrent network (CRN) is employed for complex spectral mapping where the summated power of error signals is used as the loss function for CRN training. Deep MCANC is a fixed-parameter ANC approach and large-scale multi-condition training is employed to achieve robustness against a variety of noises. We explore the performance of deep MCANC with different setups and investigate the impact of factors such as the number of loudspeakers and microphones, and the position of a secondary source, on ANC performance. Experimental results show that deep MCANC is effective for wideband noise reduction and generalizes well to untrained noises. Moreover, the proposed approach is robust against variations in reference signals and works well in the presence of nonlinear distortions.