Multi-scenario assessment of living-wall effects on building air purification, noise attenuation, and thermal insulation using quantum machine learning

Aqua Silencer: A Review of Design, Noise Attenuation, And Exhaust Emission Control in Internal Combustion Engines

ABSTRACT Supervised deep learning methods have been widely applied for seismic random noise attenuation, but their dependence on large volumes of clean training data limits their practicality. Deep image prior (DIP) provides an unsupervised alternative by exploiting the structural bias of convolutional neural networks. However, its performance is sensitive to the choice of stopping iteration and does not explicitly incorporate structural prior information inherent in seismic data. In this study, we propose a non‐local reference‐guided deep image prior framework for seismic random noise attenuation. Non‐local self‐similarity (NSS) is extended from the patch level to the pixel level to improve noise level estimation accuracy and to generate structurally consistent reference data. Based on the estimated global noise level, a noise‐driven early stopping criterion is introduced to determine the termination point of DIP optimization in a fully unsupervised manner. The NSS‐refined reference is used as the network input, allowing structural information to be incorporated into the reconstruction process. In addition, selective weight decay applied to the decoder layers further enhances the separation between signal and high‐frequency noise. Experiments on synthetic and field seismic data indicate that the proposed method effectively attenuates random noise while preserving structural continuity and reflection characteristics. Compared with existing unsupervised approaches, the method provides more stable optimization behaviour and improved reconstruction quality without requiring clean training data.

In recent years, rapid economic development and urbanization have led to serious environmental problems, including noise pollution. Although existing acoustic absorbers can be designed to target specific frequency ranges, those effective at low frequencies typically require large volumes. Consequently, there is a growing demand for compact absorbers capable of mitigating low-frequency noise in the range of 100-1500Hz. Conventional liners, such as melamine foam, exhibit limited performance in this frequency band. To address these limitations, NASA patented a broadband acoustic absorber inspired by natural reeds, which mimics their geometry and sound absorption characteristics. This study investigates the influence of reed diameter, orientation angle, and spacing on the sound absorption performance of reed-inspired structures. Models with reed spacings of 6mm, 8mm, and 10mm and reed angles of 45° were developed for reed diameters of 3.5mm and 5mm using SolidWorks. Numerical simulations were performed in COMSOL Multiphysics based on the Delany-Bazley model. The results indicate that a configuration with a 5mm reed diameter, 8mm spacing, and a 45° orientation achieves the highest sound absorption coefficient. Sound pressure measurements further confirmed the effectiveness of the optimized design, showing a reduction from 92.3dB to 48.1dB in the numerical analysis and from 91.4dB to 57.1dB in the experimental measurements, corresponding to an average sound pressure reduction of approximately 47.8%. Despite minor discrepancies between numerical and experimental results, the findings demonstrate that bio-inspired reed-like structures provide efficient and compact low-frequency noise attenuation. Overall, this study shows that careful geometric optimization of reed-inspired absorbers can bridge the gap between idealized numerical predictions and experimental performance, offering a promising approach for developing compact, high-efficiency solutions for low-frequency acoustic control in aerospace, automotive, industrial, and architectural applications.

Abstract Coupled noise suppression is one of the major denoising challenges in vertical seismic profiling (VSP) data acquired using distributed fibre-optic acoustic sensing (DAS) technology. Traditional methods to suppress coupled noise often require complicated parameter tuning, resulting in relatively low efficiency. Purely supervised deep learning methods demand a large amount of clean data as training labels, the acquisition of which is costly. Autoencoder (AE), as a deep learning framework that does not require manually annotated labels, has been widely used for random noise attenuation. In this study, we extend the AE to the suppression of coherent coupled noise in DAS-VSP data and propose a weakly supervised two-step coupled noise suppression workflow. Firstly, time windows dominated by coupled noise are manually selected from the target DAS-VSP data, and the AE is trained separately on these windows to reconstruct and remove the stable, coherent coupled noise component. Secondly, the AE with the same architecture is trained to further reconstruct the entire dataset, aiming to enhance event continuity and attenuate residual noise. We provide a mathematical argument to explain why the AE tends to reconstruct coupled noise preferentially. Both synthetic and field examples demonstrate that the proposed method can effectively suppress coupled noise in DAS-VSP data, outperforming traditional methods. Furthermore, by introducing Gaussian white noise to simulate complex noise contamination, the method is shown to be robust and generalizable to more complex noise environments. Under our experimental settings, the result further shows that for the same network depth, incorporating skip connections weakens the latent space’s preferential representation of coupled noise, leading to overall inferior performance compared with AE without skip connections.

The northern part of Paramushir Island (Kuril Arc) is dominated by the NNE-SSW oriented Vernadsky Ridge that hosts the presently active Ebeko Volcano, where phreatomagmatic eruptions coexist with complex hydrothermal activity. We apply ambient noise attenuation tomography (ANAT) to seismic data from a 21-station network, revealing three-dimensional attenuation structures. Key findings include: (1) Shallow high-attenuation anomalies ($${Q}_{c}^{-1}$$ > 0.45 at 0.5–2 km depth) beneath Ebeko’s craters and Yuriev hot springs, mapping fracture-controlled fluid reservoirs; (2) A 4–6 km deep low-attenuation core ($${Q}_{c}^{-1}$$ < 0.1) surrounded by high-attenuation zones, suggesting a magmatic storage region with radial fluid pathways; (3) A progressive north-to-south decrease in attenuation ($${Q}_{c}^{-1}$$ > 0.45 to < 0.15) that mirrors the Vernadsky ridge’s volcanic evolution from extinct southern edifices to currently active Ebeko to the north. These results illuminate how seismic attenuation patterns track both active fluid pathways and volcanic system maturation in arc environments.

Speech enhancement through denoising is essential for maintaining signal intelligibility and quality in biometric speaker verification pipelines that operate in acoustically adverse conditions. Despite the proliferation of deep learning (DL) architectures for speech denoising, simultaneously optimizing noise attenuation, perceptual fidelity, and speaker-identity preservation remains an open problem. We address this gap by benchmarking three architecturally distinct DL-based enhancement models—Wave-U-Net, CMGAN, and U-Net—on three independent, domain-diverse corpora (SpEAR, VPQAD, and Clarkson) that the models never encountered during training and by introducing commercial-grade VeriSpeak speaker-verification scores as a biometric evaluation dimension absent from prior comparative studies. Our experiments reveal a clear three-way trade-off: U-Net achieves the highest signal-to-noise ratio (SNR) gains (+61.44% on SpEAR, +67.05% on VPQAD, +235.3% on Clarkson) but sacrifices naturalness; CMGAN yields the best perceptual evaluation of speech quality (PESQ) values (3.33, 1.35, and 2.50, respectively), favoring listening-comfort applications; and Wave-U-Net delivers the strongest biometric fidelity (VeriSpeak improvements of +11.63%, +30.22%, and +29.24%) while offering competitive perceptual quality. These results highlight that model selection must be driven by the target deployment scenario and provide actionable guidance for improving biometric verification robustness under real-world noise.

Abstract This paper proposes a fractional‐order cascade extended state observer to alleviate the fundamental trade‐off between disturbance rejection and noise amplification in integer‐order observer‐based control. By incorporating fractional calculus into the cascade observer architecture, the fractional order serves as a continuous tuning parameter that captures system memory and long‐term disturbance dynamics more accurately. Frequency‐domain and stability analyses demonstrate enhanced disturbance rejection and improved noise attenuation compared with the conventional integer‐order cascade extended state observer. Numerical simulations on a 4‐degree‐of‐freedom robotic manipulator validate accurate disturbance estimation and robustness under sensor noise. The proposed approach introduces an additional degree of tuning freedom to balance performance and robustness without increasing observer structural complexity.

Simultaneous extraction of photoelastic constants and strain profiles from picosecond ultrasonic echoes via spectral analysis

Abstract Random noise attenuation remains challenging for seismic data containing complex small-scale geological structures, where conventional denoising methods often either leave residual noise or oversmooth subtle but geologically meaningful reflections. To address this issue, we propose a fx-domain quantum-adaptive basis denoising (fx-QABD) method for seismic random-noise suppression. The key idea is to model the fx-domain seismic amplitude as a signal-dependent potential and construct a Hamiltonian matrix whose eigenvectors form data-adaptive quantum basis (QAB) functions. In this representation, coherent seismic events become sparsely concentrated, whereas random noise remains broadly distributed, enabling more effective signal–noise separation through coefficient thresholding. Compared with fixed transform-based or conventional predictive filtering methods, the proposed method provides a representation that is intrinsically matched to local seismic characteristics and is therefore better suited to preserving weak and small-scale structural features. In addition, the 2D framework is extended to 3D seismic volumes through an iterative bidirectional implementation along the inline and crossline directions. Synthetic and field-data experiments demonstrate that the proposed method achieves stronger random-noise attenuation while preserving structural continuity and reflection fidelity. In the synthetic test, fx-QABD yields the highest post-denoising SNR and PSNR among the compared methods, confirming its superior capability for enhancing seismic data quality in structurally complex settings.

Abstract The literature on seismic data processing offers numerous methods for denoising single-component seismic data. However, more random noise attenuation methods explicitly for multicomponent seismic data remain underexplored. This study introduces a vector thresholding method based on the discrete orthogonal quaternion wavelet transform (DOQWT) for multicomponent seismic random noise attenuation. To address the accurate estimation of threshold value, we derive a new threshold estimation formula suitable for three-component data denoising. Furthermore, we analyze the influence of some key factors on the DOQWT vector threshold denoising performance, including threshold function, threshold value, and noise correlation. Tests on both synthetic and field ocean bottom seismometer seismic data demonstrate that the proposed DOQWT vector threshold method outperforms the conventional scalar wavelet thresholding method and concatenated vector thresholding method in terms of signal-to-noise ratio, preservation of polarization characteristics, and recovery of weak signals.

Feasibility study on noise attenuation and stability of active noise cancelling headphones for mobile hearing screening.

A ventilated silencer based on rainbow trapping structure for broadband noise attenuation

Abstract Lightweight structures with efficient vibration and noise attenuation capabilities are of great significance for modern engineering applications. Recent advancements have significantly expanded the working frequency range of topological phononic crystal plates. Nevertheless, the design of topological microstructures that enable larger areas for wave transport remains a critical challenge. This study proposes a lightweight hexagonal lattice elastic metamaterial that enables topologically protected transport of low-frequency elastic waves while maintaining structural efficiency. By tuning the geometric parameters of the composite unit cell, a double Dirac degeneracy is formed at the Γ point of the Brillouin zone, leading to a controllable topological phase transition. The band inversion between p - and d -state modes is achieved by breaking the spatial symmetry of the system, resulting in the opening of a nontrivial bandgap. A supercell composed of topologically trivial and nontrivial domains is further constructed to reveal the robust interface state in the bandgap. Further, a finite-sized metamaterial specimen was fabricated, and vibration transmission tests were also conducted to validate the numerical model. The experimental results confirm the existence of low-loss, unidirectional elastic wave propagation along the designed interface, in excellent agreement with the simulated field distributions. This study provides a new strategy for realizing lightweight and low-frequency broadband vibration isolation, while also offering a promising pathway for efficient energy harvesting.

Tunable Multi-Functional Ultra-Wide Band Gap Metamaterial with Negative Refraction for Enhanced Vibration and Noise Attenuation

With the in-depth advancement of educational informatization, classrooms, as the primary venues for teaching and learning, have increasingly attracted attention with regard to their acoustic environments. Classroom acoustics not only directly affect teaching effectiveness and students’ learning experiences, but also constitute a critical factor influencing overall instructional quality. However, in real-world teaching settings, problems such as background noise, reverberation interference, and sound propagation attenuation remain prevalent, particularly in large classrooms or specialized instructional scenarios where acoustic challenges are more pronounced. Consequently, there is a growing need to rely on intelligent speech enhancement technologies driven by deep learning algorithms to provide novel directions for upgrading and optimizing classroom acoustic environments. Against this background, this paper begins by outlining the fundamental principles of intelligent speech enhancement technology and its practical application points in classroom settings, and then focuses on examining its role in improving classroom acoustic conditions as well as its positive impact on teaching effectiveness. The findings are of significant theoretical and practical value for optimizing classroom teaching environments and promoting educational equity.

The dictionary learning approach has proven effective in seismic data denoising and interpolation. Its core advantage lies in the ability to continuously update the initial dictionary, thereby adapting to the complex structural characteristics of seismic data. However, many existing implementations rely on predefined transforms (e.g., discrete cosine transform) for dictionary initialization. These fixed, data-agnostic bases often fail to fully capture the unique features of seismic signals, which may compromise the sparsity and fidelity of signal representation. Such a limitation can significantly degrade the performance of tasks requiring high-precision reconstruction or noise attenuation. To address this issue, we propose an innovative dictionary learning framework based on a variational sparse representation model. Specifically, this framework first extracts small data patches from arbitrary locations in seismic data, and then constructs a pre-training dataset using a windowing algorithm to preserve fine-grained data features. This process yields an initial dictionary that inherently encodes the intrinsic characteristics of the input seismic data. Subsequently, the initial dictionary is separately refined and updated through the K-singular value decomposition (K-SVD) and sequential generalization of K-means (SGK) algorithms, resulting in an optimized dictionary with more accurate and data-adaptive features. In addition, we integrate a multi-iteration projections onto convex sets algorithm to compensate for missing data features, ultimately achieving high-precision seismic data interpolation and noise attenuation. Numerical experiments demonstrate that the proposed methods (variational SGK and variational K-SVD) outperform the conventional K-SVD and SGK algorithms in both interpolation accuracy and denoising performance.

The proportional integral derivative (PI) controller remains the predominant algorithm employed in engineering applications. Nevertheless, existing PI tuning methodologies, whether classical or contemporary, are often characterized by indirectness and limited accuracy or by excessive complexity that hinders practical implementation. Moreover, the influence of the noise filter incorporated within the feedback loop on the closed-loop system performance has not been comprehensively evaluated in these tuning strategies. Consequently, the resulting PI parameters frequently demonstrate suboptimal performance, necessitating empirical on-site adjustments through trial and error. To address these limitations, this study proposes a novel PI controller tuning approach that explicitly integrates the noise filter and directly designs the closed-loop system to meet specified bandwidth criteria. Additionally, the proposed method guarantees the absence of resonance peaks in the closed-loop amplitude frequency response and incorporates considerations of noise attenuation and phase margin. The efficacy and applicability of the method were validated experimentally on a permanent magnet synchronous motor (PMSM) servo system, confirming its practical utility.

Urban traffic noise poses escalating environmental challenges in rapidly urbanizing regions with high-density buildings, yet systematic investigations into its spatiotemporal characteristics remain relatively scarce. This study addresses this research gap via the synchronized on-site monitoring of traffic noise and traffic flow on a representative arterial road in Guangzhou, China. The analysis reveals that nighttime equivalent continuous A-weighted sound levels (LAeq) are 3.0–4.0 dB(A) higher than those during the congested daytime peak, a phenomenon primarily driven by higher vehicle speeds under nighttime free-flow traffic conditions. The spatial analysis uncovers complex three-dimensional noise propagation dynamics specific to urban street canyons. Vertical profiling demonstrates a counterintuitive pattern where noise levels do not attenuate with building height, and upper floors experience marginally higher noise exposure than the ground floor, which is attributed to the canyon effect, where multiple sound wave reflections offset the natural distance attenuation. A validated three-dimensional computational model was further employed to evaluate the efficacy of noise mitigation strategies, showing that an integrated intervention combining porous asphalt pavement and acoustic barriers achieves a maximum noise attenuation of 19.9 dB(A) at ground-level receptors. This significant reduction stems from a synergistic effect: porous asphalt reduces noise at the source on a global scale, while acoustic barriers provide localized shielding for the lower floors of adjacent buildings. This research concludes that effective traffic noise control in high-density urban areas requires three-dimensional, multi-faceted strategies addressing noise source characteristics, transmission pathways, and receptor vulnerabilities.

Abstract The global navigation satellite system (GNSS) is essential for timing and positioning. In conventional receivers, clock offset is treated as a common error and often lacks careful modeling. However, accurate clock state estimation is crucial in GNSS-based remote timing. Current methods typically model clock error as white noise, which can amplify estimation noise in both the up-coordinate and clock states under certain conditions. Incorporating clock modeling has the potential to mitigate such noise. This study explores the theoretical foundations of clock modeling and examines its influence on GNSS positioning and timing performance. We establish the GNSS timing model and the clock signal model, and clarify the relationship between Allan Variance and the diffusion coefficient. Using a small Rubidium atomic clock and an oven controlled crystal oscillator (OCXO) as examples, we evaluate the effect of clock modeling on frequency offset estimation noise and vertical positioning precision. Theoretical and experimental results demonstrate that clock modeling significantly reduces frequency offset estimation noise, with noise attenuation ranging from 17.19% to 52.83% for OCXO and 87.67% to 97.83% for the Rubidium clock. More stable clocks exhibit greater improvement. Additionally, clock modeling enhances short-term up-coordinate positioning stability, showing improvements of 78.74% for OCXO and 84.23% for the Rubidium clock at 1 s intervals. These findings highlight the potential of clock modeling for rapid online frequency monitoring and improved GNSS timing and positioning performance with OCXOs and compact atomic clocks.