Speech extraction from vibration signals based on deep learning.

Extracting speech information from vibration response signals is a typical system identification problem, and the traditional method is too sensitive to deviations such as model parameters, noise, boundary conditions, and position. A method was proposed to obtain speech signals by collecting vibration signals of vibroacoustic systems for deep learning training in the work. The vibroacoustic coupling finite element model was first established with the voice signal as the excitation source. The vibration acceleration signals of the vibration response point were used as the training set to extract its spectral characteristics. Training was performed by two types of networks: fully connected, and convolutional. And it is found that the Fully Connected network prediction model has faster Rate of convergence and better quality of extracted speech. The amplitude spectra of the output speech signals (network output) and the phase of the vibration signals were used to convert extracted speech signals back to the time domain during the test set. The simulation results showed that the positions of the vibration response points had little effect on the quality of speech recognition, and good speech extraction quality can be obtained. The noises of the speech signals posed a greater influence on the speech extraction quality than the noises of the vibration signals. Extracted speech quality was poor when both had large noises. This method was robust to the position deviation of vibration responses during training and testing. The smaller the structural flexibility, the better the speech extraction quality. The quality of speech extraction was reduced in a trained system as the mass of node increased in the test set, but with negligible differences. Changes in boundary conditions did not significantly affect extracted speech quality. The speech extraction model proposed in the work has good robustness to position deviations, quality deviations, and boundary conditions.

Sound Quality Performance of Orthogonal Antisymmetric Composite Laminates Embedded with SMA Wires.

Orthogonal antisymmetric composite laminates embedded with shape memory alloys (SMAs) wires have the potential to improve the sound quality of vibro-acoustics by taking advantage of the special superelasticity, temperature phase transition, and pre-strain characteristics of SMAs. In this research, space discretion and mode decoupling were employed to establish a vibro-acoustic sound quality model of SMA composite laminates. The association between the structural material parameters of SMA composite laminates and the sound quality index is then approached through methodologies. Numerical analysis was implemented to discuss the effects of SMA tensile pre-strain, SMA volume fraction, and the ratio of resin-to-graphite in the matrix on the vibro-acoustic sound quality of SMA composite laminates within a temperature environment. Subsequently, the sound quality test for SMA composite laminates is thus completed. The theoretically predicted value appears to agree well with the experimental outcomes, which validates the accuracy and applicability of the dynamic modeling theory and method for the sound quality of SMA composite laminates. The results indicate that attempting to alter the SMA tensile pre-strain, SMA volume fraction, and matrix material ratio can be used to modify loudness, sharpness, and roughness, which provides new ideas and a theoretical foundation for the design of composite laminates with decent sound quality.

Infrared Target Detection Based on Joint Spatio-Temporal Filtering and L1 Norm Regularization.

Infrared target detection is often disrupted by a complex background, resulting in a high false alarm and low target recognition. This paper proposes a robust principal component decomposition model with joint spatial and temporal filtering and L1 norm regularization to effectively suppress the complex backgrounds. The model establishes a new anisotropic Gaussian kernel diffusion function, which exploits the difference between the target and the background in the spatial domain to suppress the edge contours. Furthermore, in order to suppress the dynamically changing background, we construct an inversion model that combines temporal domain information and L1 norm regularization to globally constrain the low rank characteristics of the background, and characterize the target sparse component with L1 norm. Finally, the overlapping multiplier method is used for decomposition and reconstruction to complete the target detection.Through relevant experiments, the proposed background modeling method in this paper has a better background suppression effect in different scenes. The average values of the three evaluation indexes, SSIM, BSF and IC, are 0.986, 88.357 and 18.967, respectively. Meanwhile, the proposed detection method obtains a higher detection rate compared with other algorithms under the same false alarm rate.

Sound Transmission Loss of Metamaterial Honeycomb Core Sandwich Plate Elastically Connected with Periodic Subwavelength Arrays of Shunted Piezoelectric Patches.

Honeycomb core sandwich plates are widely used as a lightweight, high-strength sound insulation material. However, they do not perform well in specific frequency bands. Acoustic metamaterials can break the law of mass in specific frequency bands and have high sound transmission loss (STL); however, the resonance frequency is difficult to regulate. To solve this problem, this paper first proposes an infinitely large metamaterial honeycomb core sandwich plate, which can generate newly tuned piezoelectric resonance frequencies, and we study its STL. The structure has piezoelectric patches arranged in sub-wavelength arrays with inductance shunting circuits that are elastically connected to both sides of the honeycomb core sandwich plate. The effective dynamic mass density and effective dynamic bending stiffness of the metamaterial plates were obtained using the effective medium (EM) method. A theoretical model for the numerical calculation of oblique STL and diffuse-field STL was established by the structural bending wave method. The finite element simulation method was used to verify that the metamaterial plates can generate three peaks at 1147 Hz, 1481 Hz and 1849 Hz in oblique or diffuse-field STL curves, which reached 57 dB, 86 dB and 63 dB, respectively, and are significantly better than the plate rigidly connected with piezoelectric sheets and the bare plate with the same mass. In order to better understand the characteristics of STL, the explicit functions of the resonance frequencies were derived. Key influencing factors were analyzed, and the regulation law of new piezoelectric resonance frequencies was clarified.

Acoustic Insulation Mechanism of Membrane-Type Acoustic Metamaterials Loaded with Arbitrarily Shaped Mass Blocks of Variable Surface Density.

Membrane-type acoustic metamaterials (MAMs) have recently received widespread attention due to their good low-frequency sound-transmission-loss (STL) performance. A fast prediction method for the STL of rectangular membranes loaded with masses of arbitrary shapes and surface density values is proposed as a semi-analytical model for calculating the STL of membrane-type acoustic metamaterials. Through conformal mapping theory, the mass blocks of arbitrary shapes were transformed into regular shapes, which simplified the calculation model of acoustic propagation loss prediction, and the prediction results were verified by finite element simulations. The results show that the change in mass surface density was closely related to the size and frequency distribution of STL. The influence of the mass center on the STL and characteristic frequency of the film metamaterial is discussed.

Modal Performance of Two-Fiber Orthogonal Gradient Composite Laminates Embedded with SMA.

A gradient composite laminate that was composed of two­phase fibers, a shape memory alloy (SMA), and graphite was prepared to investigate modal performance and improve vibration behavior. The stress­strain relation of the single­layer composite plates was derived from Kirchhoff thin plate theory and the material constitutive of the SMA. A gradient distribution model and the eigenvalue equations of gradient composite laminates were developed. The influence of the fiber component content gradient distribution, pre­strain, the two­phase fiber volume fraction, and geometric parameters on the modal performance was analyzed. This study provides a method to avoid the structural resonance of composite laminates that are embedded with an SMA through the gradient distribution of two­phase fiber content that leads to the interaction of the material properties.