Efficient stochastic parallel gradient descent training for on-chip optical processor

In recent years,space-division multiplexing(SDM)technology,which involves transmitting data information on multiple parallel channels for efficient capacity scaling,has been widely used in fiber and free-space optical communication sys-tems.To enable flexible data management and cope with the mixing between different channels,the integrated reconfig-urable optical processor is used for optical switching and mitigating the channel crosstalk.However,efficient online train-ing becomes intricate and challenging,particularly when dealing with a significant number of channels.Here we use the stochastic parallel gradient descent(SPGD)algorithm to configure the integrated optical processor,which has less com-putation than the traditional gradient descent(GD)algorithm.We design and fabricate a 6×6 on-chip optical processor on silicon platform to implement optical switching and descrambling assisted by the online training with the SPDG algorithm.Moreover,we apply the on-chip processor configured by the SPGD algorithm to optical communications for optical switching and efficiently mitigating the channel crosstalk in SDM systems.In comparison with the traditional GD al-gorithm,it is found that the SPGD algorithm features better performance especially when the scale of matrix is large,which means it has the potential to optimize large-scale optical matrix computation acceleration chips.

A reweighted damped singular spectrum analysis method for robust seismic noise suppression

(Multichannel)Singular spectrum analysis is considered as one of the most effective methods for seismic incoherent noise suppression.It utilizes the low-rank feature of seismic signal and regards the noise suppression as a low-rank reconstruction problem.However,in some cases the seismic geophones receive some erratic disturbances and the amplitudes are dramatically larger than other receivers.The presence of this kind of noise,called erratic noise,makes singular spectrum analysis(SSA)reconstruction unstable and has undesirable effects on the final results.We robustify the low-rank reconstruction of seismic data by a reweighted damped SSA(RD-SSA)method.It incorporates the damped SSA,an improved version of SSA,into a reweighted framework.The damping operator is used to weaken the artificial disturbance introduced by the low-rank projection of both erratic and random noise.The central idea of the RD-SSA method is to iteratively approximate the observed data with the quadratic norm for the first iteration and the Tukeys bisquare norm for the rest iterations.The RD-SSA method can suppress seismic incoherent noise and keep the reconstruction process robust to the erratic disturbance.The feasibility of RD-SSA is validated via both synthetic and field data examples.

DPT‐tracker:Dual pooling transformer for efficient visual tracking

Transformer tracking always takes paired template and search images as encoder input and conduct feature extraction and target‐search feature correlation by self and/or cross attention operations,thus the model complexity will grow quadratically with the number of input images.To alleviate the burden of this tracking paradigm and facilitate practical deployment of Transformer‐based trackers,we propose a dual pooling transformer tracking framework,dubbed as DPT,which consists of three components:a simple yet efficient spatiotemporal attention model(SAM),a mutual correlation pooling Trans-former(MCPT)and a multiscale aggregation pooling Transformer(MAPT).SAM is designed to gracefully aggregates temporal dynamics and spatial appearance information of multi‐frame templates along space‐time dimensions.MCPT aims to capture multi‐scale pooled and correlated contextual features,which is followed by MAPT that aggregates multi‐scale features into a unified feature representation for tracking prediction.DPT tracker achieves AUC score of 69.5 on LaSOT and precision score of 82.8 on Track-ingNet while maintaining a shorter sequence length of attention tokens,fewer parameters and FLOPs compared to existing state‐of‐the‐art(SOTA)Transformer tracking methods.Extensive experiments demonstrate that DPT tracker yields a strong real‐time tracking baseline with a good trade‐off between tracking performance and inference efficiency.

An artificial systems,computational experiments and parallel execution-based surface electromyogram-driven anti-disturbance zeroing neurodynamic strategy for upper limb human-robot interaction control

In recent years,intelligent robots are extensively applied in the field of the industry and intelligent rehabilitation,wherein the human-robot interaction(HRI)control strategy is a momentous part that needs to be ameliorated.Specially,the efficacy and robustness of the HRI control algorithm in the presence of unknown external disturbances deserve to be addressed.To deal with these urgent issues,in this study,artificial systems,computational experiments and a parallel execution intelligent control framework are constructed for the HRI control.The upper limb-robotic exoskeleton system is re-modelled as an artificial system.Depending on surface electromyogram-based subject's active motion intention in the practical system,a non-convex function activated anti-disturbance zeroing neurodynamic(NC-ADZND)controller is devised in the artificial system for parallel interaction and HRI control with the practical system.Furthermore,the linear activation function-based zeroing neurodynamic(LAF-ZND)controller and proportionalderivative(posterior deltoid(PD))controller are presented and compared.Theoretical results substantiate the global convergence and robustness of the proposed controller in the presence of different external disturbances.In addition,the simulation results verify that the NC-ADZND controller is better than the LAF-ZND and the PD controllers in respect of convergence order and anti-disturbance characteristics.

Transfer force perception skills to robot‐assisted laminectomy via imitation learning from human demonstrations

A comparative study of two force perception skill learning approaches for robot‐assisted spinal surgery,the impedance model method and the imitation learning(IL)method,is presented.The impedance model method develops separate models for the surgeon and patient,incorporating spring‐damper and bone‐grinding models.Expert surgeons'feature parameters are collected and mapped using support vector regression and image navi-gation techniques.The imitation learning approach utilises long short‐term memory networks(LSTM)and addresses accurate data labelling challenges with custom models.Experimental results demonstrate skill recognition rates of 63.61%-74.62%for the impedance model approach,relying on manual feature extraction.Conversely,the imitation learning approach achieves a force perception recognition rate of 91.06%,outperforming the impedance model on curved bone surfaces.The findings demonstrate the potential of imitation learning to enhance skill acquisition in robot‐assisted spinal surgery by eliminating the laborious process of manual feature extraction.

DeepGCN based on variable multi‐graph and multimodal data for ASD diagnosis

Diagnosing individuals with autism spectrum disorder(ASD)accurately faces great chal-lenges in clinical practice,primarily due to the data's high heterogeneity and limited sample size.To tackle this issue,the authors constructed a deep graph convolutional network(GCN)based on variable multi‐graph and multimodal data(VMM‐DGCN)for ASD diagnosis.Firstly,the functional connectivity matrix was constructed to extract primary features.Then,the authors constructed a variable multi‐graph construction strategy to capture the multi‐scale feature representations of each subject by utilising convolutional filters with varying kernel sizes.Furthermore,the authors brought the non‐imaging in-formation into the feature representation at each scale and constructed multiple population graphs based on multimodal data by fully considering the correlation between subjects.After extracting the deeper features of population graphs using the deep GCN(DeepGCN),the authors fused the node features of multiple subgraphs to perform node classification tasks for typical control and ASD patients.The proposed algorithm was evaluated on the Autism Brain Imaging Data Exchange I(ABIDE I)dataset,achieving an accuracy of 91.62%and an area under the curve value of 95.74%.These results demon-strated its outstanding performance compared to other ASD diagnostic algorithms.

Multifunctional mixed analog/digital signal processor based on integrated photonics

Photonic signal processing offers a versatile and promising toolkit for contemporary scenarios ranging from digital optical communication to analog microwave operation.Compared to its electronic counterpart,it eliminates inherent bandwidth limitations and meanwhile exhibits the potential to provide unparalleled scalability and flexibility,particularly through integrated photonics.However,by far the on-chip solutions for optical signal processing are often tailored to specific tasks,which lacks versatility across diverse applications.Here,we propose a streamlined chip-level signal processing architecture that integrates different active and passive building blocks in silicon-on-insulator(SOI)platform with a compact and efficient manner.Comprehensive and in-depth analyses for the architecture are conducted at levels of device,system,and application.Accompanied by appropriate configuring schemes,the photonic circuitry supports loading and processing both analog and digital signals simultaneously.Three distinct tasks are facilitated with one single chip across several mainstream fields,spanning optical computing,microwave photonics,and optical communications.Notably,it has demonstrated competitive performance in functions like image processing,spectrum filtering,and electro-optical bandwidth equalization.Boasting high universality and a compact form factor,the proposed architecture is poised to be instrumental for next-generation functional fusion systems.

Dynamics and Fault Diagnosis of Railway Vehicle Gearboxes:A Review

The railway vehicle gearbox is an important part of the railway vehicle traction transmission system which ensures the smooth running of railway vehicles.However,as the running speed of railway vehicles continues to increase,the railway vehicle gearbox is exposed to a more demanding operating environment.Under both internal and external excitations,the gearbox is prone to faults such as fatigue cracks,and broken teeth.It is crucial to detect these faults before they result in severe failures and accidents.Therefore,understanding the dynamics and fault diagnosis of railway vehicle gearbox is needed.At present,there is a lack of systematic review of railway vehicle gearbox dynamics and fault diagnosis.So,this paper systematically summarizes the research progress on railway vehicle gearbox dynamics and fault diagnosis.To this end,this paper first summarizes the latest research progress on the dynamics of railway vehicle gearboxes.The dynamics and vibration characteristics of the gearbox are summarized under internal and external excitations,as well as faulty conditions.Then,the stateof-the-art signal processing and artificial intelligence methods for fault diagnosis of railway vehicle gearboxes are reviewed.In the end,future research prospects are given.

Denoising Fault-Aware Wavelet Network:A Signal Processing Informed Neural Network for Fault Diagnosis

Deep learning(DL) is progressively popular as a viable alternative to traditional signal processing(SP) based methods for fault diagnosis. However, the lack of explainability makes DL-based fault diagnosis methods difficult to be trusted and understood by industrial users. In addition, the extraction of weak fault features from signals with heavy noise is imperative in industrial applications. To address these limitations, inspired by the Filterbank-Feature-Decision methodology, we propose a new Signal Processing Informed Neural Network(SPINN) framework by embedding SP knowledge into the DL model. As one of the practical implementations for SPINN, a denoising fault-aware wavelet network(DFAWNet) is developed, which consists of fused wavelet convolution(FWConv), dynamic hard thresholding(DHT),index-based soft filtering(ISF), and a classifier. Taking advantage of wavelet transform, FWConv extracts multiscale features while learning wavelet scales and selecting important wavelet bases automatically;DHT dynamically eliminates noise-related components via point-wise hard thresholding;inspired by index-based filtering, ISF optimizes and selects optimal filters for diagnostic feature extraction. It’s worth noting that SPINN may be readily applied to different deep learning networks by simply adding filterbank and feature modules in front. Experiments results demonstrate a significant diagnostic performance improvement over other explainable or denoising deep learning networks. The corresponding code is available at https://github. com/alber tszg/DFAWn et.

Time-frequency Feature Extraction Method of the Multi-Source Shock Signal Based on Improved VMD and Bilateral Adaptive Laplace Wavelet

Vibration signals have the characteristics of multi-source strong shock coupling and strong noise interference owing to the complex structure of reciprocating machinery.Therefore,it is difficult to extract,analyze,and diagnose mechanical fault features.To accurately extract sensitive features from the strong noise interference and unsteady monitoring signals of reciprocating machinery,a study on the time-frequency feature extraction method of multi-source shock signals is conducted.Combining the characteristics of reciprocating mechanical vibration signals,a targeted optimization method considering the variational modal decomposition(VMD)mode number and second penalty factor is proposed,which completed the adaptive decomposition of coupled signals.Aiming at the bilateral asymmetric attenuation characteristics of reciprocating mechanical shock signals,a new bilateral adaptive Laplace wavelet(BALW)is established.A search strategy for wavelet local parameters of multi-shock signals is proposed using the harmony search(HS)method.A multi-source shock simulation signal is established,and actual data on the valve fault are obtained through diesel engine fault experiments.The fault recognition rate of the intake and exhaust valve clearance is above 90%and the extraction accuracy of the shock start position is improved by 10°.

A Fault Feature Extraction Model in Synchronous Generator under Stator Inter-Turn Short Circuit Based on ACMD and DEO3S

This paper proposed a new diagnosis model for the stator inter-turn short circuit fault in synchronous generators.Different from the past methods focused on the current or voltage signals to diagnose the electrical fault,the sta-tor vibration signal analysis based on ACMD(adaptive chirp mode decomposition)and DEO3S(demodulation energy operator of symmetrical differencing)was adopted to extract the fault feature.Firstly,FT(Fourier trans-form)is applied to the vibration signal to obtain the instantaneous frequency,and PE(permutation entropy)is calculated to select the proper weighting coefficients.Then,the signal is decomposed by ACMD,with the instan-taneous frequency and weighting coefficient acquired in the former step to obtain the optimal mode.Finally,DEO3S is operated to get the envelope spectrum which is able to strengthen the characteristic frequencies of the stator inter-turn short circuit fault.The study on the simulating signal and the real experiment data indicates the effectiveness of the proposed method for the stator inter-turn short circuit fault in synchronous generators.In addition,the comparison with other methods shows the superiority of the proposed model.

Compound Fault Diagnosis for Rotating Machinery:State-of-the-Art,Challenges,and Opportunities

Compound fault,as a primary failure leading to unexpected downtime of rotating machinery,dramatically increases the difficulty in fault diagnosis.To deal with the difficulty encountered in implementing compound fault diagnosis(CFD),researchers and engineers from industry and academia have made numerous significant breakthroughs in recent years.Admittedly,many systematic surveys focused on fault diagnosis have been conducted by reputable researchers.Nevertheless,previous review articles paid more attention to fault diagnosis with several single or independent faults,resulting in that there is still lacking a comprehensive survey on CFD.Therefore,to fulfill the above requirements,it is necessary to provide an in-depth overview of fault diagnosis methods or algorithms for compound faults of rotating machinery and uncover potential challenges or opportunities that would guide and inspire readers to devote their efforts to promoting fault diagnosis technology more effective and practical.Specifically,the backgrounds,including the related definitions and a new taxonomy of CFD methods,are detailed according to the way of implementing compound fault recognition.Then,the stateof-the-art applications of CFD are overviewed based on relevant publications in the past decades.Finally,the challenges and opportunities associated with implementing CFD are concluded and followed by a conclusion for ending this survey.We believe that this review article can provide a systematic guideline of CFD from different aspects for potential readers and seasoned researchers.

Comparison of uniform resampling and nonuniform sampling direct-reconstruction methods in k-space for FD-OCT

The nonuniform distribution of interference spectrum in wavenumber k-space is a key issue to limit the imaging quality of Fourier-domain optical coherence tomography(FD-OCT).At present,the reconstruction quality at different depths among a variety of processing methods in k-space is still uncertain.Using simulated and experimental interference spectra at different depths,the effects of common six processing methods including uniform resampling(linear interpolation(LI),cubic spline interpolation(CSI),time-domain interpolation(TDI),and K-B window convolution)and nonuniform sampling direct-reconstruction(Lomb periodogram(LP)and nonuniform discrete Fourier transform(NDFT))on the reconstruction quality of FD-OCT were quantitatively analyzed and compared in this work.The results obtained by using simulated and experimental data were coincident.From the experimental results,the averaged peak intensity,axial resolution,and signal-to-noise ratio(SNR)of NDFT at depth from 0.5 to 3.0mm were improved by about 1.9 dB,1.4 times,and 11.8 dB,respectively,compared to the averaged indices of all the uniform resampling methods at all depths.Similarly,the improvements of the above three indices of LP were 2.0 dB,1.4 times,and 11.7 dB,respectively.The analysis method and the results obtained in this work are helpful to select an appropriate processing method in k-space,so as to improve the imaging quality of FD-OCT.

A Review on Intelligent Detection and Classification of Power Quality Disturbances:Trends,Methodologies,and Prospects

With increasing global concerns about clean energy in smart grids,the detection of power quality disturbances(PQDs)caused by energy instability is becoming more and more prominent.It is well acknowledged that the PQD effects on power grid equipment are destructive and hazardous,which causes irreversible damage to underlying electrical/electronic equipment of the concerned intelligent grids.In order to ensure safe and reliable equipment implementation,appropriate PQDdetection technologiesmust be adopted to avoid such adverse effects.This paper summarizes the newly proposed and traditional PQD detection techniques in order to give a quick start to new researchers in the related field,where specific scenarios and events for which each technique is applicable are also clearly presented.Finally,comments on the future evolution of PQD detection techniques are given.Unlike the published review articles,this paper focuses on the new techniques from the last five years while providing a brief recap on traditional PQD detection techniques so as to supply researchers with a systematic and state-of-the-art review for PQD detection.

An Efficient Text-Independent Speaker Identification Using Feature Fusion and Transformer Model

Automatic Speaker Identification(ASI)involves the process of distinguishing an audio stream associated with numerous speakers’utterances.Some common aspects,such as the framework difference,overlapping of different sound events,and the presence of various sound sources during recording,make the ASI task much more complicated and complex.This research proposes a deep learning model to improve the accuracy of the ASI system and reduce the model training time under limited computation resources.In this research,the performance of the transformer model is investigated.Seven audio features,chromagram,Mel-spectrogram,tonnetz,Mel-Frequency Cepstral Coefficients(MFCCs),delta MFCCs,delta-delta MFCCs and spectral contrast,are extracted from the ELSDSR,CSTRVCTK,and Ar-DAD,datasets.The evaluation of various experiments demonstrates that the best performance was achieved by the proposed transformer model using seven audio features on all datasets.For ELSDSR,CSTRVCTK,and Ar-DAD,the highest attained accuracies are 0.99,0.97,and 0.99,respectively.The experimental results reveal that the proposed technique can achieve the best performance for ASI problems.

Underwater Noise Target Recognition Based on Sparse Adversarial Co-Training Model with Vertical Line Array

The automatic identification of underwater noncooperative targets without label records remains an arduous task considering the marine noise interference and the shortage of labeled samples.In particular,the data-driven mechanism of deep learning cannot identify false samples,aggravating the difficulty in noncooperative underwater target recognition.A semi-supervised ensemble framework based on vertical line array fusion and the sparse adversarial co-training algorithm is proposed to identify noncooperative targets effectively.The sound field cross-correlation compression(SCC)feature is developed to reduce noise and computational redundancy.Starting from an incomplete dataset,a joint adversarial autoencoder is constructed to extract the sparse features with source depth sensitivity,aiming to discover the unknown underwater targets.The adversarial prediction label is converted to initialize the joint co-forest,whose evaluation function is optimized by introducing adaptive confidence.The experiments prove the strong denoising performance,low mean square error,and high separability of SCC features.Compared with several state-of-the-art approaches,the numerical results illustrate the superiorities of the proposed method due to feature compression,secondary recognition,and decision fusion.

A Distributed Newton Method for Processing Signals Defined on the Large-Scale Networks

In the graph signal processing(GSP)framework,distributed algorithms are highly desirable in processing signals defined on large-scale networks.However,in most existing distributed algorithms,all nodes homogeneously perform the local computation,which calls for heavy computational and communication costs.Moreover,in many real-world networks,such as those with straggling nodes,the homogeneous manner may result in serious delay or even failure.To this end,we propose active network decomposition algorithms to select non-straggling nodes(normal nodes)that perform the main computation and communication across the network.To accommodate the decomposition in different kinds of networks,two different approaches are developed,one is centralized decomposition that leverages the adjacency of the network and the other is distributed decomposition that employs the indicator message transmission between neighboring nodes,which constitutes the main contribution of this paper.By incorporating the active decomposition scheme,a distributed Newton method is employed to solve the least squares problem in GSP,where the Hessian inverse is approximately evaluated by patching a series of inverses of local Hessian matrices each of which is governed by one normal node.The proposed algorithm inherits the fast convergence of the second-order algorithms while maintains low computational and communication cost.Numerical examples demonstrate the effectiveness of the proposed algorithm.

Big Data Analytics Using Graph Signal Processing

The networks are fundamental to our modern world and they appear throughout science and society.Access to a massive amount of data presents a unique opportunity to the researcher’s community.As networks grow in size the complexity increases and our ability to analyze them using the current state of the art is at severe risk of failing to keep pace.Therefore,this paper initiates a discussion on graph signal processing for large-scale data analysis.We first provide a comprehensive overview of core ideas in Graph signal processing(GSP)and their connection to conventional digital signal processing(DSP).We then summarize recent developments in developing basic GSP tools,including methods for graph filtering or graph learning,graph signal,graph Fourier transform(GFT),spectrum,graph frequency,etc.Graph filtering is a basic task that allows for isolating the contribution of individual frequencies and therefore enables the removal of noise.We then consider a graph filter as a model that helps to extend the application of GSP methods to large datasets.To show the suitability and the effeteness,we first created a noisy graph signal and then applied it to the filter.After several rounds of simulation results.We see that the filtered signal appears to be smoother and is closer to the original noise-free distance-based signal.By using this example application,we thoroughly demonstrated that graph filtration is efficient for big data analytics.

Automatic depression recognition by intelligent speech signal processing:A systematic survey

Depression has become one of the most common mental illnesses in the world.For better prediction and diagnosis,methods of automatic depression recognition based on speech signal are constantly proposed and updated,with a transition from the early traditional methods based on hand‐crafted features to the application of architectures of deep learning.This paper systematically and precisely outlines the most prominent and up‐to‐date research of automatic depression recognition by intelligent speech signal processing so far.Furthermore,methods for acoustic feature extraction,algorithms for classification and regression,as well as end to end deep models are investigated and analysed.Finally,general trends are summarised and key unresolved issues are identified to be considered in future studies of automatic speech depression recognition.

Implementation of VLSI on Signal Processing-Based Digital Architecture Using AES Algorithm

Continuous improvements in very-large-scale integration(VLSI)technology and design software have significantly broadened the scope of digital signal processing(DSP)applications.The use of application-specific integrated circuits(ASICs)and programmable digital signal processors for many DSP applications have changed,even though new system implementations based on reconfigurable computing are becoming more complex.Adaptable platforms that combine hardware and software programmability efficiency are rapidly maturing with discrete wavelet transformation(DWT)and sophisticated computerized design techniques,which are much needed in today’s modern world.New research and commercial efforts to sustain power optimization,cost savings,and improved runtime effectiveness have been initiated as initial reconfigurable technologies have emerged.Hence,in this paper,it is proposed that theDWTmethod can be implemented on a fieldprogrammable gate array in a digital architecture(FPGA-DA).We examined the effects of quantization on DWTperformance in classification problems to demonstrate its reliability concerning fixed-point math implementations.The Advanced Encryption Standard(AES)algorithm for DWT learning used in this architecture is less responsive to resampling errors than the previously proposed solution in the literature using the artificial neural networks(ANN)method.By reducing hardware area by 57%,the proposed system has a higher throughput rate of 88.72%,reliability analysis of 95.5%compared to the other standard methods.