An augmented diffusion algorithm with bidirectional communication for a distributed active noise control system

Recent studies on diffusion adaptation for distributed active noise control (DANC) systems have attracted significant research interest due to their balance between computational burden and stability compared to conventional centralized and decentralized adaptation schemes. The conventional multitask diffusion FxLMS algorithm assumes that the converged solutions of all control filters are consistent to each other, which is unrealistic in practice hence results in inferior performance in noise reduction. An augmented diffusion FxLMS algorithm has been proposed to overcome this problem, which adopts a neighborhood-wide adaptation and node-based combination approach to mitigate the bias in the converged solution of the multitask diffusion algorithms. However, the improvement comes at the expense of a higher computational burden and communication cost. All existing DANC systems, including the multitask and augmented diffusion algorithms, assume one-way communication between nodes. By contrast, this paper proposes a bidirectional communication scheme for the augmented diffusion algorithm to further reduce the memory requirement, computational burden, and communication cost. Simulation results in the free field and with measured room impulse responses both demonstrate that the proposed augmented diffusion algorithm with bidirectional communication can achieve a faster convergence speed than that based on one-way communication with a lower memory, computation, and communication burden.

Low-Latency Active Noise Control Using Attentive Recurrent Network

Processing latency is a critical issue for active noise control (ANC) due to the causality constraint of ANC systems. This paper addresses low-latency ANC in the context of deep learning (i.e. deep ANC). A time-domain method using an attentive recurrent network (ARN) is employed to perform deep ANC with smaller frame sizes, thus reducing algorithmic latency of deep ANC. In addition, we introduce a delay-compensated training to perform ANC using predicted noise for several milliseconds. Moreover, a revised overlap-add method is utilized during signal resynthesis to avoid the latency introduced due to overlaps between neighboring time frames. Experimental results show the effectiveness of the proposed strategies for achieving low-latency deep ANC. Combining the proposed strategies is capable of yielding zero, even negative, algorithmic latency without affecting ANC performance much, thus alleviating the causality constraint in ANC design.

Some Practical Acoustic Design and Typical Control Strategies for Multichannel Active Noise Control

Active noise control (ANC) systems usually involve a large number of loudspeakers and error microphones in order to achieve noise reduction over an extended region of space. Although fundamentals of ANC theory and principles of ANC methods have been well-established over the past 40 years, applications of this technology are facing new challenges. A larger quiet zone with better noise reduction performance is always desirable in a variety of real-life scenarios. This paper presents several important factors that affect the performance of multichannel ANC systems in some popular applications such as windows with natural ventilation and quiet-zone around heads. The factors affecting acoustic design include the reflection of a baffle plate, arrangement of error sensors in open areas, and so on. In addition, different control strategies are compared and analyzed, including centralized, decentralized, and distributed strategies. All these strategies are discussed from the signal processing side, which should be considered after a proper acoustic design. One of the important aims of this paper is to provide practical guidance for acoustic design and discuss several typical control strategies for multichannel ANC systems.

Summarisation, Simulation and Comparison of Nine Control Algorithms for an Active Control Mount with an Oscillating Coil Actuator

With the further development of the automotive industry, the traditional vibration isolation method is difficult to meet the requirements for wide frequency bands under multiple operating conditions, the active control mount (ACM) is gradually paid attentions, and the control algorithm plays a decisive role. In this paper, the ACM with oscillating coil actuator (OCA) is taken as the object, and the comparative study of the control algorithms is performed to select the optimal one for ACM. Through the modelling of ACM, the design of controller and the system simulations, the force transmission rate is used to compare the vibration isolation performance of the nine control algorithms, which are least mean square (LMS) adaptive feedforward control, recursive least square (RLS) adaptive feedforward control, filtered reference signal LMS (FxLMS) adaptive control, linear quadratic regulator (LQR) optimal control, H2 control, H∞ control, proportional integral derivative (PID) feedback control, fuzzy control and fuzzy PID control. In summary, the FxLMS adaptive control algorithm has the better performance and the advantage of easier hardware implementation, and it can apply in ACMs.

Deep ANC: A deep learning approach to active noise control

Traditional active noise control (ANC) methods are based on adaptive signal processing with the least mean square algorithm as the foundation. They are linear systems and do not perform satisfactorily in the presence of nonlinear distortions. In this paper, we formulate ANC as a supervised learning problem and propose a deep learning approach, called deep ANC, to address the nonlinear ANC problem. The main idea is to employ deep learning to encode the optimal control parameters corresponding to different noises and environments. A convolutional recurrent network (CRN) is trained to estimate the real and imaginary spectrograms of the canceling signal from the reference signal so that the corresponding anti-noise can eliminate or attenuate the primary noise in the ANC system. Large-scale multi-condition training is employed to achieve good generalization and robustness against a variety of noises. The deep ANC method can be trained to achieve active noise cancellation no matter whether the reference signal is noise or noisy speech. In addition, a delay-compensated strategy is introduced to solve the potential latency problem of ANC systems. Experimental results show that deep ANC is effective for wideband noise reduction and generalizes well to untrained noises. Moreover, the proposed method can achieve ANC within a quiet zone and is robust against variations in reference signals.

Performance Evaluation of Active Noise Control for a Real Device Casing

Reduction of noise generated by devices is an important problem, both in industrial environments where high sound pressure levels may damage hearing, and in households where the sound pressure level is usually moderate, but may cause discomfort and stress. Classically used passive methods often have poor performance for low frequencies. Alternatively, active control can be used to improve noise reduction in this frequency range. In the proposed approach, noise generated by devices may be reduced by controlling vibrations of the casing. The authors previously confirmed the performance of the proposed active control approach using a dedicated noise-canceling casing. Herein, we describe further development and application of the method to an off-the-shelf washing machine. Electrodynamic actuators were installed on four walls of the washing machine. The performance of the control systems was experimentally evaluated during the real spinning phase and the results are reported here.

The application of a multi-reference control strategy to noise cancelling headphones

Active noise cancelling (ANC) headphones have seen significant commercial success and a number of control strategies have been proposed, including feedforward, feedback, and hybrid configurations, using both analogue and digital implementations. Irrespective of the configuration or implementation approach, the strategies proposed in the open-literature have focused on implementations where the control system for each ear of the headphones operates independently. In this paper, a multi-reference ANC strategy is proposed and investigated for noise cancelling headphones. As with standard feedforward ANC headphones, the system utilises a single error microphone and single reference microphone on each cup; however, in the proposed configuration, the left and right reference microphones are used to achieve control at both the left and right ear cups. The performance of this controller design is compared to a standard single reference feedforward controller implementation under a variety of different sound field conditions. Although the proposed strategy requires an increased computational demand, it is shown that there is a significant control advantage for noise sources originating from the side of the user, whilst the performance for front and rear sources is maintained.

Active Structural Acoustic Control of an Active Casing Placed in a Corner

Electric appliances used in workplaces and everyday life often generate a low-frequency noise, which affects human body systems. Passive methods employed to reduce noise are not effective at low frequencies. The classical approach to active noise control practically involves the generation of local zones of quiet, whereas at other areas the noise is reinforced. Moreover, it usually requires a large number of secondary sound sources. Hence, an active casing approach has been developed. The active casing panels’ vibrations are controlled to reduce the device noise emission. Efficiency of this method has been previously confirmed by the authors and the results have been reported in multiple journal publications. However, in the previous research experiments, the active casing was placed at a distance from the enclosure walls. In this research, the active casing is located in a corner and such placement is intentionally used to facilitate the active control system’s operation. The noise reduction performance is investigated at multiple configurations, including a range of distances from the corner and different error microphone arrangements. The analysis of both primary and secondary paths is given. Advantages and drawbacks of different active casing configurations are presented and discussed.

Active control of outgoing noise fields in rooms

Current active noise control systems can cancel noises in a duct effectively. However, they are insufficient for suppressing complex noise fields in time-varying rooms. This paper develops an active noise control system that can cancel tonal noise fields produced by a primary source in a room. The problem of tonal noise field control is formulated as estimating and canceling the outgoing field on a sphere surrounding the primary source. The proposed system limits the energy of the primary source radiating out of the sphere, thereby creating a global quiet zone inside the room. In addition, it removes the need for online secondary path estimation with reduced influence on desired sound fields in the room. A method for estimating the outgoing field on a sphere is presented, together with a wave-domain algorithm for controlling the outgoing field. Simulations and hardware demonstrations show the proposed system can reduce tonal noise fields in a room and over a wide frequency range.

Performance evaluation of an active headrest considering non-stationary broadband disturbances and head movement

An application of active noise control (ANC) is the active headrest, using actuators (loadspeakers), sensors (microphones), and a controller to create a local zone of quiet around the occupants head. In this paper, the attenuation performance of various ANC-algorithms for active headrests known from literature is evaluated considering non-stationary broadband disturbances and head movement. Numerical studies are performed to determine the optimal plant models and parameters for the ANC-algorithms. Based on the findings of the numerical studies, several real-time experiments are conducted with and without head tracking examining the distribution of the 10 dB zone of quiet and the attenuation at the occupants ear using either a head mounted microphone technique, the remote microphone technique, the virtual microphone technique, or the virtual microphone control method. It is found that none of the algorithms using a virtual sensing technique can produce a 10 dB zone of quiet for all considered non-stationary broadband disturbances and head movement. For the algorithm using a head mounted microphone, it is possible to form a 10 dB zone of quiet, but placing a microphone at the ear is not feasible in most situations.

Feasibility study of transfer function model on electrocardiogram change caused by acupuncture

Background

Acupuncture treatments that regulate the heart are used to treat various clinical disorders and conditions. Although many studies have been conducted to measure quantitatively the effects of acupuncture, thus far, models that describe these effects have not been established. The purpose of this study was to derive a transfer function model of acupuncture stimulation within the electrocardiograms based on the periods before, during, and after acupuncture.

Methods

Fourteen healthy subjects were included in this clinical trial. Five-minute electrocardiograms were captured before, during, and after acupuncture at HT7. For each period, signal-averaged electrocardiograms were created from all of the subjects’ 5-min electrocardiograms for that period. Individual transfer functions, which has the highest average goodness of fit, were derived for each period pair. By averaging individual transfer functions, generalized transfer functions were derived.

Results

The transfer function with the highest average goodness of fit was a fraction with 4th order numerator and 5th order denominator. Fourteen individual transfer functions were derived separately for each pair of periods: before and during acupuncture, during and after acupuncture, and before and after acupuncture. Three generalized transfer functions were derived by averaging individual transfer functions for each period pair.

Conclusion

The three generalized transfer functions that were derived may reflect the electrocardiogram changes caused by acupuncture. However, this clinical trial included only 14 subjects. Further studies with control groups and more subjects are needed.

This clinical trial has been registered on the Clinical Research Information Service, Republic of Korea (No. KCT0001944). The first enrolment of subject started at 2 June 2015 and this trial was retrospectively registered at 14 June 2016

Multichannel active noise control for spatially sparse noise fields

Multi-channel active noise control (ANC) is currently an attractive solution for the attenuation of low-frequency noise fields, in three-dimensional space. This paper develops a controller for the case when the noise source components are sparsely distributed in space. The anti-noise signals are designed as in conventional ANC to minimize the residual errors but with an additional term containing an ℓ_l norm regularization applied to the signal magnitude. This results in that only secondary sources close to the noise sources are required to be active for cancellation of sparse noise fields. Adaptive algorithms with low computational complexity and faster convergence speeds are proposed.

Potential-based methodology for active sound control in three dimensional settings

This paper extends a potential-based approach to active noise shielding with preservation of wanted sound in three-dimensional settings. The approach, which was described in a previous publication [Lim et al., J. Acoust. Soc. Am. 129(2), 717-725 (2011)], provides several significant advantages over conventional noise control methods. Most significantly, the methodology does not require any information including the characterization of sources, impedance boundary conditions and surrounding medium, and that the methodology automatically differentiates between the wanted and unwanted sound components. The previous publication proved the concept in one-dimensional conditions. In this paper, the approach for more realistic conditions is studied by numerical simulation and experimental validation in three-dimensional cases. The results provide a guideline to the implementation of the active shielding method with practical three-dimensional conditions. Through numerical simulation it is demonstrated that while leaving the wanted sound unchanged, the developed approach offers selective volumetric noise cancellation within a targeted domain. In addition, the method is implemented in a three-dimensional experiment with a white noise source in a semi-anechoic chamber. The experimental study identifies practical difficulties and limitations in the use of the approach for real applications.

Modeling, analysis, and validation of an active T-shaped noise barrier

With ever-increasing land traffic, abatement of traffic noise using noise barriers remains significant, yet it is a challenging task due to spatial competition with other infrastructure. In this study, a deep insight into the diffraction characteristics of acoustic fields near noise barriers of various geometries and surface conditions was achieved using numerical simulations. A T-shaped passive noise barrier with acoustically soft upper surfaces was demonstrated to outperform other candidates in a middle- or high-frequency range. Based on attributes of the acoustic field diffracted by T-shaped barriers, an active control strategy was developed to revamp the T-shaped barrier, in which a filtered minimax algorithm was established to drive the secondary sound sources. This algorithm resulted in more uniformly distributed residual sound fields than a filtered-X least mean square algorithm. Performance of the actively controlled barrier was evaluated at different positions and spacings of secondary sound sources and error sensors, leading to a series of optimal criteria for the design of active noise barriers. A prototype was fabricated and validated experimentally, manifesting particular effectiveness in insulating low-frequency noise, supplementing well the capacity of a passive T-shaped barrier which is effective in the middle- or high-frequency range.

Nonlinear Active Noise Control via Model-Based Approaches

Active noise control has attracted much research attention due to its several advantages over passive noise control. This paper introduces two model-based noise canceling techniques, that is, using the Moving Average (MA) model and a feedforward Neural Network (NN) to estimate the signal. The Least Mean Square (LMS) algorithm is used to minimize the error in the MA model while a backpropagation algorithm is employed to optimize the NN. Due to its advantages of good robustness and nonlinear processing, the NN is considered to be suitable for nonlinear signals. In order to reduce computational cost, the backpropagation algorithm in the NN is applied once at each time step with only one iteration. To examine the methods, two real-world problems are considered, one being engine noise and the other road traffic noise. A comparison between the two methods is carried out. Results indicate that both the MA and NN processors are effective in reducing the noises and that the NN based approach is superior over the MA model, especially for low frequency band.

A model for three-dimensional simulation of acoustic emissions from rotating machine vibration

Industrial noise can be successfully mitigated with the combined use of passive and active noise control (ANC) strategies. In a noisy area, a practical solution for noise attenuation may include both the use of baffles and ANC. When the operator is required to stay in movement in a delimited spatial area, conventional ANC is usually not able to adequately cancel the noise over the whole area. Control strategies need to be devised to achieve acceptable spatial coverage. A three-dimensional vibration-acoustic model is proposed in this paper. The signal of an accelerometer attached to the bulk of a centrifugal pump installed in an empty room was used as the input of this model. The signal of a microphone that changes its position in a spatial grid inside this room is the output. In the first stage, the ARX models are used to describe a SISO system where the input is the machine vibration and the output is the noise level measured at a certain point. In the second stage, spatial interpolation is used to estimate the model parameters. Results show good agreement between experimental data and model predictions, indicating the potential of using the model for the design of ANC.

Improved training of neural networks for the nonlinear active control of sound and vibration

Active control of sound and vibration has been the subject of a lot of research in recent years, and examples of applications are now numerous. However, few practical implementations of nonlinear active controllers have been realized. Nonlinear active controllers may be required in cases where the actuators used in active control systems exhibit nonlinear characteristics, or in cases when the structure to be controlled exhibits a nonlinear behavior. A multilayer perceptron neural-network based control structure was previously introduced as a nonlinear active controller, with a training algorithm based on an extended backpropagation scheme. This paper introduces new heuristical training algorithms for the same neural-network control structure. The objective is to develop new algorithms with faster convergence speed (by using nonlinear recursive-least-squares algorithms) and/or lower computational loads (by using an alternative approach to compute the instantaneous gradient of the cost function). Experimental results of active sound control using a nonlinear actuator with linear and nonlinear controllers are presented. The results show that some of the new algorithms can greatly improve the learning rate of the neural-network control structure, and that for the considered experimental setup a neural-network controller can outperform linear controllers.

A stability analysis of a decentralized adaptive feedback active control system of sinusoidal sound in free space

In some cases, the implementation of active control of sound in free space requires a large number of secondary sources and error sensors. In terms of control hardware, this may translate into considerable processing power requirement. A practical method to decrease processing power is to decentralize the control; that is, implement many single-input, single-output independent controllers operating simultaneously instead of a large multiple-input, multiple-output system. The main drawback of decentralized control is the risk of global instability. The purpose of this paper is to derive conditions under which globally stable control system behavior can be obtained in the case of adaptive feedback decentralized control for a sinusoidal disturbance. The main objective is to give practical conditions derived from the small gain theorem and the Nyquist criterion for the stability of the control system. These conditions only take into account the geometrical arrangement of the secondary sources and error sensors. This analysis involves a new parameter beta called "performance index," which is associated with both the convergence of the individual controllers and the global stability of the system. Simulation and experimental results are shown to illustrate the effectiveness of the developed analytical tools.

Spatially extended sound equalization in rectangular rooms

The results of a theoretical study on global sound equalization in rectangular rooms at low frequencies are presented. The zone where sound equalization can be obtained is a continuous three-dimensional region that occupies almost the complete volume of the room. It is proved that the equalization of broadband signals can be achieved by the simulation of a traveling plane wave using FIR filters. The optimal solution has been calculated following the traditional least-squares approximation, where a modeling delay has been applied to minimize reverberation. An advantage of the method is that the sound field can be estimated with sensors placed in the limits of the equalization zone. As a consequence, a free space for the listeners can be obtained.

Broadband control of plate radiation using a piezoelectric, double-amplifier active-skin and structural acoustic sensing

The potential of a piezoelectric, double-amplifier active-skin with structural acoustic sensing (SAS) is demonstrated for the reduction of broadband acoustic radiation from a clamped, aluminum plate. The active-skin is a continuous covering of the vibrating portions of the plate with active, independently controllable piezoelectric, double-amplifier elements and is designed to affect control by altering the continuous structural radiation impedance rather than structural vibration. In simulation, acoustic models are sought for the primary and secondary sources that incorporate finite element methods. Simulation indicates that a total radiated power attenuation in excess of 10 dB may be achieved between 250 and 750 Hz with microphone error sensing, while under SAS the radiated power is reduced by nearly 8 dB in the same frequency range. In experiment, the adaptive feed forward filtered-x LMS (least mean square) algorithm, implemented on a Texas Instruments C40 DSP, was used in conjunction with the 6I6O control system. With microphone error sensing, 11.8-dB attenuation was achieved in the overall radiated power between 175 and 600 Hz, while inclusion of SAS resulted in a 7.3-dB overall power reduction in this frequency band.

Active cancellation system of acoustic noise in MR imaging

In this paper, we introduce a new neural-network architecture for reducing the acoustic noise level in magnetic resonance (MR) imaging processes. The proposed neural network (NN) consists of two cascaded time-delay NN's (TDNN's). This NN is used as the predictor of a feedback active noise control (ANC) system for reducing acoustic noises. Experimental results with real MR noises show that the proposed system achieved an average noise power attenuation of 18.75 dB, which compares favorably with previous studies. Preliminary results also show that with the proposed ANC system installed, acoustic MR noises are greatly attenuated while verbal communication during MRI sessions is not affected.

An adaptive noise reduction stethoscope for auscultation in high noise environments

Auscultation of lung sounds in patient transport vehicles such as an ambulance or aircraft is unachievable because of high ambient noise levels. Aircraft noise levels of 90-100 dB SPL are common, while lung sounds have been measured in the 22-30 dB SPL range in free space and 65-70 dB SPL within a stethoscope coupler. Also, the bandwidth of lung sounds and vehicle noise typically has significant overlap, limiting the utility of traditional band-pass filtering. In this study, a passively shielded stethoscope coupler that contains one microphone to measure the (noise-corrupted) lung sound and another to measure the ambient noise was constructed. Lung sound measurements were made on a healthy subject in a simulated USAF C-130 aircraft environment within an acoustic chamber at noise levels ranging from 80 to 100 dB SPL. Adaptive filtering schemes using a least-mean-squares (LMS) and a normalized least-mean-squares (NLMS) approach were employed to extract the lung sounds from the noise-corrupted signal. Approximately 15 dB of noise reduction over the 100-600 Hz frequency range was achieved with the LMS algorithm, with the more complex NLMS algorithm providing faster convergence and up to 5 dB of additional noise reduction. These findings indicate that a combination of active and passive noise reduction can be used to measure lung sounds in high noise environments.

Active control of vibration using a neural network

Feedforward control of sound and vibration using a neural network-based control system is considered, with the aim being to derive an architecture/algorithm combination which is capable of supplanting the commonly used finite impulse response filter/filtered-x least mean square (LMS) linear arrangement for certain nonlinear problems. An adaptive algorithm is derived which enables stable adaptation of the neural controller for this purpose, while providing the capacity to maintain causality within the control scheme. The algorithm is shown to be simply a generalization of the linear filtered-x LMS algorithm. Experiments are undertaken which demonstrate the utility of the proposed arrangement, showing that it performs as well as a linear control system for a linear control problem and better for a nonlinear control problem. The experiments also lead to the conclusion that more work is required to improve the predictability and consistency of the performance before the neural network controller becomes a practical alternative to the current linear feedforward systems.