Piezoelectric actuators (PAs) require high precision positioning for the applications of micro electrical mechanical systems, but it exhibits hysteresis nonlinearity which deteriorates positioning accuracy if no proper compensation is given. Hysteresis nonlinear modeling of PAs is a prime choice for hysteresis compensation. This paper proposes a novel intelligent positioning control algorithm based on Bouc-Wen (BW) model for the compensation of a bi-morph type piezoelectric actuator (PA) suffering rate-dependent hysteresis. A region based mixed-species swarm optimization (RMSO) algorithm is proposed for BW modeling to capture the dynamic nonlinearity of a piezoelectric actuator which exhibits rate-dependent hysteresis. Results of numerical simulations have been disclosed to illustrate the performance enhancement of RMSO over classical algorithm while they are applied to the parameter fitting problem of BW model for experimentally acquired datasets. An model based adaptive Fuzzy neural network (Fuzzy-NN) controller of PA is utilized to compensate the hysteresis for the positioning tracking control. Experimental results also illustrate the good performance of the proposed RMSO-BW based control scheme for the hysteresis compensation control of the PA.

Objective . Sleep staging is the basis for sleep quality assessment and diagnosis of sleep-related disorders. In response to the inadequacy of traditional manual judgement of sleep stages, using machine learning techniques for automatic sleep staging has become a hot topic. To improve the performance of sleep staging, numerous studies have extracted a large number of sleep-related characteristics. However, there are redundant and irrelevant features in the high-dimensional features that reduce the classification accuracy. To address this issue, an effective hybrid feature selection method based on the entropy weight method is proposed in this paper for automatic sleep staging. Approach . Firstly, we preprocess the four modal polysomnography (PSG) signals, including electroencephalogram (EEG), electrooculogram (EOG), electrocardiogram (ECG) and electromyogram (EMG). Secondly, the time domain, frequency domain and nonlinear features are extracted from the preprocessed signals, with a total of 185 features. Then, in order to acquire characteristics of the multi-modal signals that are highly correlated with the sleep stages, the proposed hybrid feature selection method is applied to choose effective features. This method is divided into two stages. In stage I, the entropy weight method is employed to combine two filter methods to build a subset of features. This stage evaluates features based on information theory and distance metrics, which can quickly obtain a subset of features and retain the relevant features. In stage II, Sequential Forward Selection is used to evaluate the subset of features and eliminate redundant features. Further more, to achieve better performance of classification, an ensemble model based on support vector machine, K-nearest neighbor, random forest and multilayer perceptron is finally constructed for classifying sleep stages. Main results . The experiment using the Cyclic Alternating Pattern (CAP) sleep database is performed to assess the performance of the method proposed in this paper. The proposed hybrid feature selection method chooses only 30 features highly correlated to sleep stages. The accuracy, F1 score and Kappa coefficient of 6 class sleep staging reach 88.86%, 83.15% and 0.8531%, respectively. Significance . Experimental results show the effectiveness of the proposed method compared to the existing state-of-the-art studies. It greatly reduces the number of features required while achieving outstanding auto-sleep staging results.

Using traditional FMECA to analyze the hazard of subway sliding plug door system, there are problems such as easy-to-take repetitive values, irrational allocation of expert's weights, and failure to consider the weights of evaluation factors. Hence, this paper proposes to improve FMECA by using linear interpolation to increase the differentiation of the same fault probability occurrence among various fault modes. Apply the DUOWA to dynamically assign weights to different experts. The AHP is used to endow weights to diverse evaluation factors to make them more suitable for engineering needs. The operation records of metro trains were collected from the Shanghai subway manufacturing plant for 1836 days. The hazards of the 17 common faults of the door were first analyzed using an improved FMECA,the EDCU is the component most prone to failure. Next, use a RCM strategy to determine maintenance periods for different faults. Finally, through the Weibull distribution fitting test, the fault rate function of the door is obtained, and the RUL of the door is predicted.

Speaker recognition is an indispensable technology for biometrics. It distinguishes individuals based on their vocal patterns. In this paper, a joint confirmation method based on the Akaike Information Criterion (AIC) of reconstruction error (REE) and time complexity (AIC-Time joint confirmation method) is proposed to select the optimal decomposition rank of NMF. Furthermore, non-negative matrix factorization (NMF) is applied to the spectrogram to generate speaker features. The network for speaker recognition is based on Convolutional Neural Networks combining Squeeze Excitation (SE) modules with ResNeXt, and the best combination is explored experimentally. The SE module conducts a channel-level adaptive adjustment of the feature maps, reducing redundancy and noise interference while improving feature extraction efficiency and accuracy. The ResNeXt convolutional neural network concurrently executes multiple convolutional kernels, acquiring richer feature information. The experimental results demonstrate that compared to speaker recognition based on Gaussian mixture models (GMM), Visual Geometry Group Network (VGGNet), ResNet, and SE-ResNeXt using spectrograms, this method increases the accuracy by an average of 5.8% and 16.24% under the overlaid of babble and factory1 noise with different signal-to-noise ratios, respectively. In the short speech test, the test set is short speech of 1s and 2s, and the noise is superimposed. Compared with other methods, the recognition rate is increased by an average of 8.67% and 11.72%, respectively.

Tram vehicles share roadway resources with other traffic participants, creating a significant risk of collision. Therefore, obstacle-detection technology is crucial for guaranteeing safe driving. Traditional obstacle ranging methods based on picture geometry sometimes rely on the road’s vanishing point or the known size of the obstruction to determine the camera’s pitch angle. Nevertheless, when the measurements of the unstructured road or obstruction are ambiguous, these methods often fail to determine the camera pitch angle. This research introduces a monocular ranging method derived from the geometric characteristics of the track. This study presents a monocular ranging method derived from the geometric characteristics of the track. First, a tram track detection model based on geometric moments is constructed, extracting track edge information from the color space and utilizing geometric moments to extract track features to obtain the starting point information of the track curve. Second, the camera pitch angle is determined by computing the lateral distance variation of the track, utilizing the starting point coordinates of the track curve and fixed track spacing. Finally, the YOLO approach is employed for obstacle identification, and the distance to obstacles is quantified using a developed obstacle range model grounded in image geometry. Static and dynamic ranging experiments were performed to assess the efficacy and resilience of the proposed ranging algorithm. In the static ranging experiment, the approach attained average errors of 1.32% for longitudinal ranging, 5.15% for lateral ranging, and 1.26% for straight-line ranging. Among these, the average errors in the longitudinal and straight-line ranges surpassed those of current techniques. In the dynamic ranging experiment, the average error in the straight-line range was 2.72%, further substantiating the effectiveness of this technique in practical applications.

A fixed split in the second heart sound ( S ₂ ) indicates an aortic septal defect (ASD). This work aims at the shortcomings of using the time interval ( T A₂→P₂ ) between the sounds produced by the aortic valve closure ( A ₂ ), as well as the pulmonary valve closure ( P ₂ ) to evaluate the fixed split S ₂ , which is T A₂→P₂ easily influenced by the heartbeat and difficult to distinguish from other types of splits S ₂ without considering the entire respiratory phase and the third heart sound ( S ₃ ). Thus, this study proposes a novel methodology for detecting an ASD using estimated split coefficients of S ₂ ( S _2split ) combined with the number ( N S₂ ) of S ₂ covered by the moving window with 9 s. The important contributions are highlighted as follows: 1) A novel, simple and efficient methodology based on the Gaussianmixturemodel (GMM) is proposed for estimating the sounds of A ₂ and P ₂ . 2) An overlapping rate ( OLR )-based estimated S _2split is proposed to overcome the split degree influenced by the heartbeat when using the parameter T A₂→P₂ . 3) To achieve high precision of the wide splitting of ASDs, the statistics of N S₂ - S _2split and N S₂ - T A₂→P₂ are proposed for analysis. 4) To avoid drawing a unilateral conclusion, the correct ratio ( CR ) is proposed to predict an ASD. The criteria, derived based on the combination of μ NS₂,S_2split + σ NS₂,S_2split ⩾ 0.45, μ NS₂,S_2split ⩾ 0.6, σ NS₂,S_2split ⩽ 0.01, 0.05 ⩽ μ NS₂,TA₂→P₂ ⩽ 0.1, and CR ⩾ 0.9, are determined to detect ASD. The performance evaluation, as opposed to the state-of-the-art algorithms, demonstrates that the overall accuracy, F1 score and Cohens kappa coefficient achieved by the proposed algorithm all exceed up to 3% compared to the second-highest performance metrics.