Shift Invariance Research Articles

Robust Gridless Estimation of Angles and Delays for Full-Dimensional Wideband mmWave Channels

In this paper, we consider robust channel estimation for a millimeter wave (mmWave) massive MIMO system with uniform planar arrays (UPA). For many gridless angle estimation methods of mmWave channels, the channel gains needs to be time-invariant during training. We propose a gridless method that is applicable to time-invariant and time-varying channels, and the proposed method is robust to channel variations. This is done by employing time-varying precoders and combiners that are banded block Toeplitz (BBT) matrices. The advantage of time-varying BBT precoders is two-fold. First, with time-varying BBT precoding we can fabricate time-varying channels. Thus the channel can be estimated as if it is a time-varying channel whether the actual channel is time-varying or not. We will show that, thanks to fabricated time variation in the channel, a small training overhead can be used. Secondly, BBT precoding and combining preserve the shift invariance property of the received data matrix. Due to the preservation of shift invariance property, we can employ most existing direction finding methods for estimating the angles and delays, including low complexity gridless subspace based algorithms. Simulation results are given to show that the proposed time-varying BBT precoding provides robust and accurate channel estimates with a low complexity.

An Enhancement Process for Multi-focus Images Resulted from Image Fusion using qshiftN DTCWT and MPCA in Multiresolution Domain

Images having different focus values are combined using image fusion methodologies. The fused image shows superiority in the quality and highly informative than compared o any other multifocus images. The fused image would show its suitability for visual perception object detection due to its high quality than compared to multifocus images. Among the various quality parameters of the fused image, directional invariance and shift invariance significantly influence the fused image quality. The traditional image fusion methods using wavelets severely suffers with poor invariance and lack of directionality. An efficient image fusion method is developed using DTCWT with qshiftN and MPCA algorithms in the multiresolution (MR) domain. Multifocus input images are decomposed into high and low frequency components using MR algorithm. The decomposed frequency components of the input images are fused using DTCWT with qshiftN algorithm. The proposed fusion algorithm preserves both sift invariance and directional properties of the multifocus source image. Lastly, the fused image is processed using MPCA algorithm to enhance the features. The proposed methodology is assessed using various multifocus images and the evaluated metrics are contrasted with technologically advanced algorithms reported recently. The statistical metrics evaluated for the proposed method for different multifocus images shows superior properties than compared to the recently reported multifocus image fusion algorithms.

An ESPRIT-Based Supervised Channel Estimation Method Using Tensor Train Decomposition for mmWave 3-D MIMO-OFDM Systems

This paper focuses on the downlink supervised channel estimation problem for the millimeter wave threedimensional multiple-input multiple-output orthogonal frequency division multiplexing (mmWave 3D MIMO-OFDM) systems, where both the transmitter and the receiver are equipped with uniform rectangular arrays (URAs). Based on the sparse scattering nature, the mmWave channel is modeled as a low-rank higher-order tensor. By formulating the channel and the received training signal as tensors, a fast ESPRIT-based Vandermondestructured tensor decomposition method is proposed to estimate the channel parameters involving angles of arrival and departure (AoAs/AoDs), delays and path gains. In specific, we first establish the relationship of the tensor ranks between the tensor train (TT) model and the higher-order tensor CANDECOMP/PARAFAC (CP) signal model. Then the TT decomposition is used to estimate the signal subspace and derive the shift invariance equation, which exploits the higher-order tensor low-rankness of the signal and the Vandermonde structure in the frequency domain. Based on the derived analytical estimation errors, theoretical justifications are provided to unveil the advantage of using TT decomposition. Moreover, we extend the proposed method in an iterative manner to pursue higher accuracy. According to our simulation results, compared with the state-of-the-art, the proposed method achieves higher estimation accuracy on both the channel parameters and the entire channel. In addition, up to 87% of computing time can be saved in comparison to the current best iterative algorithm in the experiments.

Proposal of terrestrial-device wake-up signal transmission method with Doppler variation pre-compensation on IoT platform via LEO satellite

Regarding the waking up of terrestrial Internet of Things devices via a low Earth orbit satellite downlink signal, it is difficult to demodulate the wake-up signal received by such devices that are directly below a satellite due to Doppler variation caused by the satellite’s high-speed movement. This letter proposes a device wake-up signal transmission method for reducing the effect of Doppler shift variation on a device’s receive signal by adding frequency variation to the wake-up signals that cancels out the Doppler variation and transmitting the signal in downlink. Computer simulations showed that the proposed method enables the waking up of devices directly below a satellite.

Delay-induced vibrational resonance in the Rayleigh–Plesset bubble oscillator

We examine the impacts of time-delay and phase shift between two acoustic driving forces on vibrational resonance (VR) phenomena in the oscillations of a spherical gas bubble. Using the approximate method of direct separation of the motions, we obtain the equation of slow motion and the response amplitude, and we validate the theoretical predictions with numerical simulations. We find that the response amplitude of the system at the lower frequency varies periodically with respect to the phase shift. When the phase shift consists of an even number of periods, it can be optimized to enhance the system’s response in the relevant parameter space of the high-frequency driving force. In addition to the enhancement of the VR peak by variation of the phase shift, our results show that the time-delay also plays a significant role in the bubble’s response to dual-frequency acoustic driving fields. It and can be exploited either to suppress drastically, or to modulate, the resonance peaks, thereby controlling the resonances. Our analysis shows further that cooperation between the time-delay and the amplitude of the high-frequency component of the acoustic waves can induce multiple resonances. These results could potentially be exploited to control and enhance ultrasonic cleaning processes by varying the time-delay parameter in the presence of phase shifted dual-frequency acoustic waves. Moreover, it could be employed to achieve improved accuracy in ultrasonic biomedical diagnosis and tumour therapy, as well as for targeted delivery of reagents transported within bubbles.

Physical effects of geocosmic storms. Part 3

A review of the results of observing the effects of the most interesting ionospheric and magnetic storms that took place in 2018–2019 is presented. Ionospheric and magnetic storms are considered as components of geocosmic storms. In accordance with the systemic paradigm of L.F. Chernogor, a geocosmic storm is a complex of processes caused by a solar storm, reduced to synergistically interacting storms in a magnetic field (magnetic storm), ionosphere (ionospheric storm), atmosphere (atmospheric storm) and electric field atmospheric-ionospheric-magnetospheric origin (electric storm). Using the analysis of specific magnetic and ionospheric storms as an example, it is shown that, in addition to general patterns, each storm has its own individual characteristics. The expediency of a comprehensive study of the manifestations of each new geocosmic storm has been confirmed. A wide variety of ionospheric and magnetic storms have been confirmed. A weak magnetic storm is not necessarily accompanied by a weak ionospheric storm, while a strong one is accompanied by a strong one. An ionospheric storm can be single-phase (positive or negative) or multi-phase, when the phases alternate. Storms can last from several hours to many days. Storms are accompanied by both aperiodic and quasi-periodic variations. The latter are caused by atmospheric gravity waves (periods 10–180 min) and infrasound (period less than 5 min). Of all the components of a geocosmic storm, except for academic ones, ionospheric storms are of the greatest practical interest. It is these storms that limit the potential characteristics of radio systems for various purposes (telecommunications, radio navigation, radar, remote radio sounding, radio astronomy). It is shown that ionospheric storms are accompanied by the appearance of multipath, broadening of the Doppler spectra (and even their collapse), significant aperiodic and quasi-periodic variations in the Doppler frequency shift, the amplitude of radio signals, and the trajectories of radio waves in the decameter range. In particular, the height of reflection of radio waves changed during ionospheric storms by 50–180 km. The electron concentration during moderate negative ionospheric storms decreased up to 2 times, and during moderate positive ionospheric storms it increased up to 3 times. Under the action of atmospheric gravity waves, the relative amplitude of quasi-periodic variations in the electron concentration reached 50%, and due to low-frequency infrasound, it did not exceed 0.7%.

Simultaneous optimization of inventory, maintenance, and quality for production systems subject to multiple mean and variance shifts

This article presents a general, integrated production, quality, and condition-based maintenance model for imperfect processes with multiple out-of-control operating states. The production system is subject to multiple quality disruptions, which affect both the process mean and variance, and failures. The quality of the production process output is monitored by a fully adaptive control scheme, and the state of deterioration is estimated through periodic sampling inspection. The proposed control scheme’s operation is modeled through a Markov chain, and the optimal quality control, maintenance, and inventory policy are defined based on an expected total cost minimization criterion. Finally, an illustrative example is employed to demonstrate the operation of the proposed model, sensitivity analysis is conducted to generate insights into how the process parameters misspecification affects the economic performance, and a comparative study is performed to evaluate the model’s superiority against the respective single quality shift model.

The dental phenotype of primary dentition in SATB2-associated syndrome: a report of three cases and literature review

BackgroundSATB2-associated syndrome (SAS; OMIM: 612,313) is an autosomal dominant inherited multisystemic disorder caused by several variants of the SATB2 gene. SAS is characterized by intellectual disability, developmental delay, severe speech anomalies, craniofacial anomalies, and dental abnormalities. Here, we report the dental phenotype of primary dentition of three Chinese children with SAS.Case presentationAll three cases with SAS showed intellectual disability, speech and language anomalies, and palate anomalies. For the dental phenotype, all three cases showed macrodontia, crowded dentition, extensive caries, periapical abscesses and fistulas. Radiographs showed the wide-open root apex of deciduous teeth, loss of mandibular second bicuspids, delayed root formation of permanent teeth, rotated teeth, and taurodontism. Sanger sequencing of case 1 showed that there was a heterozygous code shift variation, c1985delT (p.F662Sfs*9) in the SATB2 gene, which has not been reported in literature. Root canal therapy, carious restoration, and teeth extraction were managed promptly, while preventive dental care was given regularly.ConclusionsThe dental phenotype of primary dentition in SAS may show macrodontia, crowded dentition, severe caries, wide-open root apex of deciduous teeth, loss of mandibular second bicuspids, delayed root formation of permanent teeth, rotated teeth, and taurodontism. Regular oral hygiene instructions and preventive dental care are both required.

A comparative study on X-ray peak broadening analysis of mechanically alloyed Al2O3 particles dispersion strengthened Al 7017 alloy

AA7017 + x vol. % Al2O3 (x = 5, 10, 20) nanocomposites powders were synthesized by HEBM process with milling time of 0 h and 20 h. The synthesized nanocomposite powders were characterized using XRD (Bruker USA D8 Advance, Davinci) with CuKα (1.5406 Aͦ), 2θ range (30°-90°), and Scan Speed (1°/min) for examining the crystalline nature of the powder samples. The size and shape of the powder particle were identified by HR-SEM (FEI Quanta FEG 200F) with 30 KV voltage, wide magnification range (12x-105 x) and secondary electron mode, and HR-TEM (JEOL Japan, JEM-2100 Plus) with voltage (200 kV), and magnification range (below 50 nm). Powder particles were measured by laser scattering technique using zeta analyzer (Laser Particle size Analyser, M/s Microtrac),and finally, the melting point and peak shift variation of composite powder samples were analyzed using TG-DSC (Netzsch TG-DSC) with temperature range (25–1400 °C), and heating rate (10 °C/min)under argon atmosphere. Furthermore, the particle sizes evaluated from different models of diffraction analysis (Scherrer, Williamson-Hall, and size-strain plot, SSP, methods) were compared with the results from PSA, HR-SEM, and HR-TEM analyses. The obtained mechanical properties explained that 20h milled AA7017 + x vol. % Al2O3 (x = 5, 10, and 20) samples produced higher values; and the properties improvement in all the samples were correlated theoretically by various strengthening mechanisms.

Chemical shift variations in common metabolites

The reliability and robustness of metabolite assignments in 1H NMR is complicated by numerous factors including variations in temperature, pH, buffer choice, ionic strength, and mixture composition that led to peak overlap and spectral crowding. As sample conditions fluctuate, peak drift and line broadening further complicate peak deconvolution and subsequent chemical assignment. We present a collection of 1D 1H NMR spectra of 54 common metabolites at varied pH (6.0 to 8.0 in 0.5 step increments) and temperature (290 K to 308 K) to quantify chemical shift variability to facilitate automated metabolite assignments. Our results illustrate the fundamental challenges with accurately assigning NMR peaks under varied environmental conditions prevalent in complex mixtures. Phosphorylated metabolites showed a larger variation in chemical shifts due to pH, whereas; amino acids showed a higher variation due to temperature. Mixtures of phosphorous compounds showed a consistently poor reliability in achieving an accurate assignment. Phosphorylated cholines, amino acids, and glycerols yielded a 40 % false negative rate for 7 out of 9 mixture conditions. Amino acids had a false negative rate of 57 % at 298 K and pH 8. Our results demonstrate that the automated assignments of complex biofluid mixtures require an expert to intervene to confirm the accuracy of metabolite assignments. Our analysis also indicates the need for reference databases to include spectra under a variety of conditions that includes mixtures and a range of pH and temperature to improve the accuracy and reproducibility of metabolite assignments.

Bias variance Toeplitz Matrix based Shift Invariance classifier for building detection from satellite images

Buildings are a predominant category in man made objects of satellite images. As far as issues like building planning and shape of the building in satellite images plays a paramount aspect. Also, with the contemporary and swift deployment of technology, images are becoming more precise. Moreover, when the image correctness, the distinction in the composition and its features is increasingly obvious hence building feature identification with the aid of artificial selection has become a complicated task. Also, as the building edges are frequently adjoining to other features, there results in misclassification between final building extraction results and original buildings. Hence, novelty in building detection is of considerable importance to enhance the building detection with precise results with minimum classification (i.e., false positive rate). In this work an ideal classification method called, Toeplitz Matrix Convolutional Neural Network-based Shift Invariance (TMCNN-SI) for precise building detection from satellite images is proposed. The TMCNN-SI method is divided into three layers, namely, input layer, three hidden layers and output layer. The input layer by employing the Bias Variance Decomposition-based Generalization function obtains the feature point selection for further processing. In the first hidden layer, a Convolutional Kernel function to the convolved features is applied to generate feature map. In the second hidden layer, Toeplitz Matrix-based Shift Invariance function is applied with the objective of reducing the dimensionality involved in the generated feature map. In the third hidden layer, a Log Fusion function is applied for binary classification with minimum loss. Finally, the resultant target class is obtained in the output layer. The outline of the proposed TMCNN-SI method is studied in terms of classification time, classification accuracy and false positive rate. The radical results show that the TMCNN-SI approach bestows efficiently when considered with respect to classification.

A Hierarchical Approach to Intelligent Mission Planning for Heterogeneous Fleets of Autonomous Underwater Vehicles

The rapid development of marine robotic technology in recent decades has resulted in significant improvements in the self-sufficiency of autonomous underwater vehicles (AUVs). However, simple scenario-based approaches are no longer sufficient when it comes to ensuring the efficient interaction of multiple autonomous vehicles in complex dynamic missions. The necessity to respond cooperatively to constant changes under severe operating constraints, such as energy or communication limitations, results in the challenge of developing intelligent adaptive approaches for planning and organizing group activities. The current study presents a novel hierarchical approach to the group control system designed for large heterogeneous fleets of AUVs. The high-level core of the approach is rendezvous-based mission planning and is aimed to effectively decompose the mission, ensure regular communication, and schedule AUVs recharging activities. The high-level planning problem is formulated as an original acyclic variation of the inverse shift scheduling problem, which is NP-hard. Since regular schedule adjustments are supposed to be made by the robots themselves right in the course of the mission, a meta-heuristic hybrid evolutionary algorithm is developed to construct feasible sub-optimal solutions in a short time. The high efficiency of the proposed approach is shown through a series of computational experiments.

Ultrahigh-sensitivity label-free single mode-tapered multimode-single mode fiber U-shaped biosensor for Staphylococcus aureus detection

The rapid detection of Staphylococcus aureus (S. aureus), particularly in the case of very low levels of S. aureus, is very important in food safety and medical diagnostics. In this paper, a fast (<20 mins), label-free biosensor using a U-shaped single mode- tapered multimode- single mode (STMS) fiber structure is presented. The average wavelength shift of the sensor (modified with the IgG from porcine serum with a concentration 200 μg/mL) is as high as 1.552 nm when it is immersed into an S. aureus sample with a concentration of 4 cells/mL. By further diluting the S. aureus to 0.4 cells/mL for a sample volume of 1.5 mL, it is found that a dip wavelength shift of 0.348 nm is observed on occasion which confirms that the biosensor can detect a single colony (cell) of S. aureus. Results for the detection of S. aureus in milk and lettuce samples are also presented and are compared with the wavelength shift results in phosphate buffered saline (PBS), and it is found that the variation in the average wavelength shift is − 21% for S. aureus concentrations from 4 to 4 × 103 cells/mL.

Synthesis, crystal structure and NMR-study new mononuclear paramagnetic Er (III) complex based on imine derivatives of thiacalix[4]arene

The synthesis, crystal structure and NMR-study of new paramagnetic Er (III) complex based on imine derivatives of thiacalix[4]arene was carried out. Temperature dependences of the 1H NMR spectra of complex Er(III) and calix[4]arene with N-contained substituents (I) have been analyzed using the line shape analysis, taking into account the temperature variation of paramagnetic chemical shifts, within the frame of the dynamic NMR method. Two kinds of molecular dynamics in I are reported for the CDCl3 solution. The first one conformational dynamics is conditioned by the pinched cone A ↔ pinched cone B interconversion with activation free energy ΔG≠(298 K) = 40 ± 3 kJ/mol (observed at low temperature). The second one dynamics is connected with rearrangements in coordination environment (with ΔG≠(298 K) = 58 ± 3 kJ/mol). Due to substantial temperature dependence of paramagnetic shifts, the complex I can be considered as NMR thermosensor reagent for local temperature monitoring. Owing to the large magnetic moments typical for many Ln-ions and especially high value of magnetic anisotropy for Er cation (III) adopting capped trigonal prismatic geometry of coordination sphere one may expect that 1-Er can potentially exhibit SMM behavior.

Psychophysical Effects on an Interference Pattern in a Double-Slit Optical System

Objective: A two-year online experiment tested the hypothesis that focused human attention alternatively directed toward or away from a double-slit optical system would affect the interference pattern in a predictable, unidirectional fashion. A control condition was employed by having a web server periodically simulate a human observer. Method: Based on the results of an independent reanalysis of these data and the outcome of an independent conceptual replication, we revisited the original directional hypothesis to explore the possibility that mind-wandering and other distractions might have caused attention or intention to unpredictably fluctuate. That in turn might have caused the hypothesized psychophysical influence to be more readily detected as a bidirectional effect (i.e., a shift in variance) rather than as unidirectional effect (a shift in mean). To test this idea, we developed a variance-based analysis using data collected during the first year of the experiment and applied it to data from the second year. Results: The first year’s data showed that experimental sessions conducted by humans resulted in significant variance differences as compared to control sessions conducted by a computer, z = 4.16, p = .00002. The same analysis applied to the second year’s data resulted in z = 3.14, p = .0008. Examination of environmental and apparatus variables indicated that those factors were not responsible for the observed changes in variance. Conclusion: The results suggest that a variance analysis may be more sensitive to psychophysical effects in this type of experiment.

When disturbances favour species adapted to stressful soils: grazing may benefit soil specialists in gypsum plant communities

Background Herbivory and extreme soils are drivers of plant evolution. Adaptation to extreme soils often implies substrate-specific traits, and resistance to herbivory involves tolerance or avoidance mechanisms. However, little research has been done on the effect of grazing on plant communities rich in edaphic endemics growing on extreme soils. A widespread study case is gypsum drylands, where livestock grazing often prevails. Despite their limiting conditions, gypsum soils host a unique and highly specialised flora, identified as a conservation priority. Methods We evaluated the effect of different grazing intensities on the assembly of perennial plant communities growing on gypsum soils. We considered the contribution of species gypsum affinity and key functional traits of species such as traits related to gypsum specialisation (leaf S accumulation) or traits related to plant tolerance to herbivory such as leaf C and N concentrations. The effect of grazing intensity on plant community indices (i.e., richness, diversity, community weighted-means (CWM) and functional diversity (FD) indices for each trait) were modelled using Generalised Linear Mixed Models (GLMM). We analysed the relative contribution of interspecific trait variation and intraspecific trait variation (ITV) in shifts of community index values. Results Livestock grazing may benefit gypsum plant specialists during community assembly, as species with high gypsum affinity, and high leaf S contents, were more likely to assemble in the most grazed plots. Grazing also promoted species with traits related to herbivory tolerance, as species with a rapid-growth strategy (high leaf N, low leaf C) were promoted under high grazing conditions. Species that ultimately formed gypsum plant communities had sufficient functional variability among individuals to cope with different grazing intensities, as intraspecific variability was the main component of species assembly for CWM values. Conclusions The positive effects of grazing on plant communities in gypsum soils indicate that livestock may be a key tool for the conservation of these edaphic endemics.

Seasonal variation modifies the spatial patterns of soil microbial community structure and enzyme activity in a meadow steppe

Understanding how spatial heterogeneity in soil microbial community structure and enzyme activities vary seasonally, and identifying the underlying mechanisms, are crucial for predicting how soil organic matter dynamics and nutrient cycling may respond to environmental change. We examined spatial variability in microbial community structure and enzyme activity and measured associated changes in plant biomass, soil microclimate, soil nutrients and other soil characteristics for 75 sampling points across a 15 × 15 m area of a Leymus chinensis meadow steppe in northeastern China over three seasons (summer, fall and spring). We observed seasonal variation in the spatial patterns of soil microbial community structure and enzyme activity, and the plant and soil variables associated with this variation also changed over time. Specifically, the microbial community structure exhibited strong spatial dependence, with high a relative spatial variance [C / (C + C0) ≥ 0.9] in summer, whereas it was more homogeneous in spring and fall. In summer, the spatial patterns of plants, soil dissolved inorganic N and dissolved organic C were the most significant predictors of spatial variation in microbial community structure. In fall, a mix of plant and soil physical explanatory variables explained 22–46 % of the spatial variation in microbial community structure. Meanwhile, only soil physical variables (soil bulk density and electrical conductivity) were strongly associated with spatial variation in microbial community structure in spring. The spatial links between microbial community structure and enzyme activity were reshaped throughout the growing season. We observed strong spatial links between soil microbial community structure and enzyme activity in spring and fall. In summer, the association between soil enzyme activities and microbial biomass was decoupled, indicating that microbial biomass showed higher turnover than soil enzymes at this time. Overall, our results reveal the drivers of spatial variation in soil microbial community structure and enzyme activity shift seasonally, highlighting the value of seasonal sampling to accurately estimate the heterogeneity and complexity of ecosystem-level processes in grasslands.

NMR studies of Sugammadex formulations complexes with steroidal neuromuscular blockers drugs Rocuronium and Vecuronium

Sugammadex a modified γ-cyclodextrin forms strong water-soluble complexes with steroid neuromuscular blockers (NMBs) such as Rocuronium and Vecuronium, creating a concentration gradient that favors the displacement of these drugs from the neuromuscular junction to the plasma, quickly reversing the neuromuscular obstruction. This work described the comparative evaluation between two formulations of Sugammadex solution from Blau Farmacêutica S/A compared to reference formulation Bridion®, by Nuclear Magnetic Resonance (NMR) 1H and DOSY techniques. The formulations were evaluated in presence of Rocuronium and Vecuronium in the stoichiometric proportions of 1:1 and 1:0.4 for both NMBs. NMR 1H spectra of the formulations Blau and Bridion® did not show significant variations of chemical shifts showing similar physicochemical parameters. Chemical shifts and diffusion coefficients of NMBs were influenced by the presence of Sugammadex when compared to NMBs formulations in the absence of any cyclodextrin, directly demonstrating the strong intermolecular interaction associated with the host molecule in a high proportion. No significant variation in the diffusion coefficients of the key components between formulations in the presence of NMBs in the equimolar 1:1 and 1:04 w/w ratios was observed. The constants Ka and KD were determined and presented 16.7 and 0.06 mM values respectively, according to literature for this system.

Deep Image Inpainting With Enhanced Normalization and Contextual Attention

Deep learning-based image inpainting has been widely studied, leading to great success. However, many methods adopt convolution and normalization operations, which will bring up some issues to affect the performance. The vanilla normalization cannot distinguish the pixels in corrupted regions from the other valid pixels, resulting in the mean and variance shifts. In addition, the limited receptive field of convolution makes it unable to capture long-range valid information directly. In order to tackle these challenges, we propose a novel deep generative model for image inpainting with two key modules, namely, the channel and spatially adaptive batch normalization (CSA-BN) module, and the selective latent-space-mapping-based contextual attention (SLSM-CA) layer. We replace the vanilla normalization with the CSA-BN module. By channel and spatially adaptive denormalization, the CSA-BN module can mitigate the spatial mean and variance shifts in each channel in a targeted way. In addition, we also integrate the SLSM-CA layer into our model to capture the long-range correlations explicitly. By introducing dual-branch attention and a feature selection module, the SLSM-CA layer can selectively utilize the multi-scale background information to improve prediction quality. What’s more, it introduces the latent spaces to achieve the low-rank approximations of attention matrices and to reduce computational costs. Extensive quantitative and qualitative evaluations demonstrate the superiority of the proposed method compared with state-of-the-art methods.

Enhanced Sparse Low-Rank Representation via Nonconvex Regularization for Rotating Machinery Early Fault Feature Extraction

The weak fault feature extraction is key to early fault diagnosis of rotary machinery. However, the existing sparse low-rank fault feature extraction methods have the deficiency of underestimation and low peak signal-to-noise ratio. To solve these problems, this article presents an enhanced sparse low-rank (ESL) representation approach for weak fault feature extraction. Considering the periodic self-similarity and shift invariance of fault feature, a weighted dual approximation regularization is proposed for noise and fault irrelevant harmonic suppression, which provides a cornerstone for rotating machinery weak fault feature extraction. To be specific, the truncated nuclear norm (TNN) and weighted generalized minimax-concave (WGMC) penalty are leveraged to form the weighted dual approximation regularization so that it can inherit their superior properties. The TNN can capture the periodic self-similarity and shift-invariance structure of fault impulses while restraining the noise; the WGMC penalty can enhance the sparsity, restrain the fault irrelevant harmonics, and overcome the deficiency of underestimating the large amplitude components. Therefore, the proposed model can effectively extract the weak fault feature. The proposed approach is applied to bearing and planet gear fault diagnosis to evaluate its effectiveness. Comparison results show the significant improvements of the proposed ESL method, indicating that it has great potentials in fault diagnosis of rotating machinery.