In this article, a novel prediction index is constructed, hybrid filtering proposed, and remaining useful life (RUL) framework developed. the proposed framework, different models are built for operation states of rolling bearings. normal state, linear model built, Kalman filter (KF) implemented to determine failure start time (FST). degradation dimensionless CRRMS based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) wavelet threshold. Then, double exponential...

Classification of coronary artery stenosis is essential in assisting physicians diagnosing cardiovascular diseases. However, due to the complexity medical diagnosis and confidentiality images, it difficult obtain many image samples for scientific research general. In addition, degree, location, morphology different patients, as well noise CT angiography (CTA) make challenging extract typing features effectively. To address above problems, firstly, a joint segmentation method proposed based...

In this paper, a fault diagnosis method is proposed based on multi-sensor fusion information for single and composite of train braking systems. Firstly, the mass model brake established operating environment. Then, pre-allocation linear-weighted summation criterion are to fuse monitoring data. Finally, improved expectation maximization, modes parameters identified, faults diagnosed in real time. The simulation results show that systems can be effectively accurately identification results....

By analyzing the mechanism of pure air emergency brake for high-speed train, discrete model is established. Aiming at problem that time-varying hidden parameters cannot be observed directly, sliding window-based expectation maximization proposed, and unobserved are identified. Firstly, position size window selected; then, used parameter identification; finally, combined with gradient optimization, optimal identifications obtained. The simulation results show can identified quickly accurately...

To solve the particle degradation and sample impoverishment, a maximum correntropy-based extended filter method is proposed by optimizing important density function. In method, importance sampling guided measurement information, entropy criterion high-order moments information of signal are fully utilized, non-Gaussian heavy-tailed distribution noise with outliers effectively suppressed. The simulation results demonstrate effectiveness feasibility filtering method.

In this paper, we propose an output feedback flight tracking control scheme for helicopter attitude and altitude systems with unmeasured states under full state constraints. Firstly, a observer is constructed based on the measured signals, which proven to be rigorous since all are constrained within desired assigned scopes. Secondly, controller built using estimations constraints method. Then, Barrier Lyapunov function method adopted guarantee stability of composite closed-loop nonlinear...

In this article, a condition monitoring approach is proposed based on vibration signal, aiming at improving the adaptability of feature extraction and accuracy classification. First, original signal acquired under certain working preprocessed by dividing it into multiple segments, followed decomposition. Then, features each decomposed are extracted theory diversity entropy (DE). Two parameters in DE optimized considering fact that these crucial for classification result. The optimization...

In this paper, a state estimation method is proposed based on the multi-rate asynchronous sensors fusion with missing measurements for high-speed train. Firstly, considering actual operating environment, speed measurement space model established. Then, maximum likelihood evaluation criterion constructed, and multiple imputation recovery of intermittent continuous data. Finally, improved adaptive attenuation particle filter-unscented Kalman filter (PF-UKF) matrix weighted fusion, global fused...

In this paper, the relationship between traction, resistance, braking force, speed and acceleration is discussed. A nonlinear parametric state space model established to describe dynamic characteristics of running process high train. Further, a filtering method based on Gaussian Sum theory extended Kalman filter proposed for estimating states parameters systems, applied train no matter it affected by or non-Gaussian noise. Firstly, probability density function (PDF) stochastic noise...

Abstract Robot visual place recognition is a research hotspot, and robot systems based on 3D point clouds are even more focused. In recent years, scan-context descriptor imagery has become standard method for cloud positioning. Based its system characteristics, we integrated it with the Shuffle Attention (SA) mechanism module to improve performance. And tested public database NCLT dataset (North Campus Long-Term Vision dataset), our good results.

<title>Abstract</title> Visual place recognition using 3D lidar in robotics is a popular research topic. Feature representation descriptors, such as scan-context which are 2D descriptors obtained from point clouds, one of the directions with great in-depth value. This paper proposes new method to improve robot visual based on descriptors. Our approach focuses combination loss functions and optimizers accuracy robustness system. Specifically, we introduce custom function that encourages...

Considering the competitive failure that exists in operation of industrial systems, including degradation and sudden failure, this paper presents a reliability assessment method based on three-parameter Weibull distribution Wiener process. Nonlinear functions are proposed to establish relationship between different processes, model is derived. The Metropolis-Hastings (MH) sampling algorithm Monte Carlo Markov Chain (MCMC) employed estimate parameters study. results obtained by real samples....

The study shows that the leakage of expansion joints in underground heating primary pipe network will lead to fluctuation soil temperature and conductivity around point. Based on online monitoring changes physical quantities joint, this paper established assessment model as input parameters degree output parameters, which took radial basis function kernel made use PSO algorithm achieve optimization parameters. Then, were predicted based ARIMA model, LS-SVM joint was realized according...

In this paper, the mechanism of pure air emergency braking for high speed train is analyzed. Considering influence actual running environment on performance, an model based established. The expectation maximization identification sliding window proposed, and unobserved time varying adhesion coefficient identified. Firstly, position size are determined. Then identified by window. Finally, combined with gradient optimization, optimal obtained. simulation results show that online proposed in...

Classification detection of coronary artery stenosis plays a vital role in assisting physicians to diagnose cardiovascular diseases. To classify the degree accurately and efficiently, CT angiography (CTA) images are converted into binary by joint segmentation method, which is based on maximum between-class variance (OTSU) region growing. Further, binarized datasets normalized, realizes preprocessing medical datasets. Then, enhance classification performance, model stenosis, ResNet18 transfer...

Track irregularity detection is crucial to assure passenger comfort and operation security of the train. In this paper, a novel track estimation model based on Temporal Convolution Network (TCN) Long Short-Term Memory (LSTM) proposed monitor state promptly accurately. Firstly, denoising method combining Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) grey theory established realize noise reduction train attitude data poor signal-to-noise ratio (SNR). Then,...

Data fusion has become an important technology in all fields to cope with the current big data Era. Current data-level research algorithms have different degrees of flaws. The main purpose this paper is solve sparsity and loss information caused by L1 norm distance measurement variational Bayesian inference model fusion. Based on L2 second-order penalty, a proposed. input updated twice combining empirical module entropy values. Total Variation (TV) penalty factor applied keep signal...

As an indispensable component in the operation of machines, rolling bearings play important role modern industrial systems. In order to accurately diagnose target domain fault data this context, paper, variable working condition dataset is firstly constructed. Further, discrete Fourier transform (DFT) applied original vibration signals, which can obtain frequency signals. Then, on basis adversarial neural networks(DANN) network, multi-channel convolutional long short-term memory...

In this paper, the emergency braking of high speed train is studied, and relationship between force, resistance, deceleration discussed. According to Newton's second law motion, by discretizing displacement, model established. Further, unobserved parameters are identified iterative Expectation Maximization (EM) algorithm. Firstly, conditional expectation designed E step; then, maximized M step identify parameters. The simulation results show that, when initial value within ± 45% true value,...

The precise prediction of the remaining useful life (RUL) rolling bearing is critical for arranging maintenance, improving safety and reliability machines. Among them, construction appropriate health index has a significant impact on prediction. Based complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) regularized particle filter (RPF), method predicting RUL suggested to tackle this problem. In method, vibration signal collected by sensor decomposed based CEEMDAN....