Weight Coefficient Matrix Research Articles

Multi-label feature selection based on nonlinear mapping

Feature selection is one of the important pre-processing methods for dimensionality reduction in multi-label learning tasks, which has attracted extensive attention in recent years. Most of the existing approaches transform feature data into the label space during the feature-label mapping process by assuming linear relationship between the feature and label spaces. However, the linearity assumption does not hold in most cases, especially for high-dimensional spaces. This work proposes a novel dimension reduction model for multi-label data by using nonlinear mapping (NMFS). The model introduces a point-to-point sigmoid function to describe the intrinsic relationship from data space to label space. The proposed method improves the generalization ability of feature selection by limiting the range of data transformation to the interval [0,1] which is consistent with the predicted values of the labels. The feature weight matrix is constrained by the l2,1-norm to ensure its sparsity, which forms the basis of feature selection. The variables in the NMFS model are iteratively updated using the gradient momentum optimization strategy, and a sparse weight-coefficient matrix is obtained for multi-label feature ordering. Experimental results on 14 multi-label data sets verify the effectiveness of the proposed method, and show the proposed method is superior to the most advanced multi-label feature selection methods.

Late Fusion Multiview Clustering via Min-Max Optimization.

Multiview clustering (MVC) sufficiently exploits the diverse and complementary information among different views to improve the clustering performance. As a representative algorithm of MVC, the newly proposed simple multiple kernel k -means (SimpleMKKM) algorithm takes a min-max formulation and applies a gradient descent algorithm to decrease the resultant objective function. It is empirically observed that its superiority is attributed to the novel min-max formulation and the new optimization. In this article, we propose to integrate the min-max learning paradigm adopted by SimpleMKKM into late fusion MVC (LF-MVC). This leads to a tri-level max-min-max optimization problem with respect to the perturbation matrices, weight coefficient, and clustering partition matrix. To solve this intractable max-min-max optimization problem, we design an efficient two-step alternative optimization strategy. Furthermore, we analyze the generalization clustering performance of the proposed algorithm from the theoretical perspective. Comprehensive experiments have been conducted to evaluate the proposed algorithm in terms of clustering accuracy (ACC), computation time, convergence, as well as the evolution of the learned consensus clustering matrix, clustering with different numbers of samples, and analysis of the learned kernel weight. The experimental results show that the proposed algorithm is able to significantly reduce the computation time and improve the clustering ACC when compared to several state-of-the-art LF-MVC algorithms. The code of this work is publicly released at: https://xinwangliu.github.io/Under-Review.

Improved Model Predictive Control Path Tracking Approach Based on Online Updated Algorithm with Fuzzy Control and Variable Prediction Time Domain for Autonomous Vehicles

The design of trajectory tracking controllers for smart driving cars still faces problems, such as uncertain parameters and it being time-consuming. To improve the tracking performance of the trajectory tracking controller and reduce the computation of the controller, this paper proposes an improved model predictive control (MPC) method based on fuzzy control and an online update algorithm. First, a vehicle dynamics model is constructed and a feedforward MPC controller is designed; second, a real-time updating method of the time domain parameters is proposed to replace the previous method of empirically selecting the time domain parameters; lastly, a fuzzy controller is proposed for the real-time adjustment of the weight coefficient matrix of the model predictive controller according to the lateral and heading errors of the vehicle, and a state matrix-based cosine similarity updating mechanism is developed for determining the updating nodes of the state matrix to reduce the controller computation caused by the continuous updating of the state matrix when the longitudinal vehicle speed changes. Finally, the controller is compared with the traditional model prediction controller through the co-simulation of CARSIM and MATLAB/Simulink, and the results show that the controller has great improvement in terms of tracking accuracy and controller computational load.

Path-Tracking Control of Soft-Target Vehicle Test System Based on Compensation Weight Coefficient Matrix and Adaptive Preview Time

<div>In order to enhance the path-tracking accuracy and adaptability of the electric flatbed vehicle (EFV) in the soft-target vehicle test system, an improved controller is designed based on the linear quadratic regulator (LQR) algorithm. First, the LQR feedback controller is designed based on the EFV dynamics tracking error model, and the genetic algorithm is utilized to obtain the optimal weight coefficient matrix for different speeds. Second, a weight coefficient matrix compensation strategy is proposed to address the changes in the relationship between the vehicle–road position and attitude caused by external disturbances and the state of EFV. An offline parameter table is created to reduce the computational load on the microcontroller of EFV and enhance real-time path-tracking performance. Furthermore, an adaptive preview time control strategy is added to reduce the overshooting caused by control delay. This strategy is based on road curvature and traveling speed. Finally, the simulation experiment and field experiment are carried out. The results show that the improved LQR controller exhibits high tracking accuracy during the emergency double lane changes experiment and can maintain accuracy and stability at different speeds.</div>

Exploration and Application of a Muscle Fatigue Assessment Model Based on NMF for Multi-Muscle Synergistic Movements.

Muscle fatigue significantly impacts coordination, stability, and speed in daily activities. Accurate assessment of muscle fatigue is vital for effective exercise programs, injury prevention, and sports performance enhancement. Current methods mostly focus on individual muscles and strength evaluation, overlooking overall fatigue in multi-muscle movements. This study introduces a comprehensive muscle fatigue model using non-negative matrix factorization (NMF) weighting. NMF is employed to analyze the duration multi-muscle weight coefficient matrix (DMWCM) during synergistic movements, and four electromyographic (EMG) signal features in time, frequency, and complexity domains are selected. Particle Swarm Optimization (PSO) optimizes feature weights. The DMWCM and weighted features combine to calculate the Comprehensive Muscle Fatigue Index (CMFI) for multi-muscle synergistic movements. Experimental results show that CMFI correlates with perceived exertion (RPE) and Speed Dynamic Score (SDS), confirming its accuracy and real-time tracking in assessing multi-muscle synergistic movements. This model offers a more comprehensive approach to muscle fatigue assessment, with potential benefits for exercise training, injury prevention, and sports medicine.

Analysis of PMSM Short-Circuit Detection Systems Using Transfer Learning of Deep Convolutional Networks

Abstract Modern permanent magnet synchronous motor (PMSM) diagnostic systems are now combined with advanced artificial intelligence techniques, such as deep neural networks. However, the design of such systems is mainly focussed on a selected type of damage or motor type with a limited range of rated parameters. The application of the idea of transfer learning (TL) allows the fully automatic extraction of universal fault symptoms, which can be used for various diagnostic tasks. In the research, the possibility of using the TL idea in the implementation of PMSM stator windings fault-detection systems was considered. The method is based on the characteristic symptoms of stator defects determined for another type of motor or mathematical model in the target diagnostic application of PMSM. This paper presents a comparison of PMSM motor inter-turn short circuit fault detection systems using TL of a deep convolutional network. Due to the use of direct phase current signal analysis by the convolutional neural network (CNN), it was possible to ensure high accuracy of fault detection with simultaneously short reaction time to occurring fault. The technique used was based on the use of a weight coefficient matrix of a pre-trained structure, the adaptation of which was carried out for different sources of diagnostic information.

Fast 3D gravity and magnetic modelling using midpoint quadrature and 2D FFT

To avoid the problem of the traditional methods consuming large computational resources to calculate the kernel matrix and 2D discrete convolution, we present a novel approach for 3D gravity and magnetic modelling. This method combines the midpoint quadrature method with a 2D fast Fourier transform (FFT) to calculate the gravity and magnetic anomalies with arbitrary density or magnetic susceptibility distribution. In this scheme, we apply the midpoint quadrature method to calculate the volume element of the integral. Then, the convolution of the weight coefficient matrix with density or magnetization is efficiently computed via the 2D FFT. Finally, the accuracy and efficiency of the proposed algorithm are validated by using an artificial model and a real topography model. The numerical results demonstrate that the proposed algorithm’s computation time and the memory requirement are decreased by approximately two orders of magnitude compared with the space-wavenumber domain method.

Trajectory Generation for Multiprocess Robotic Tasks Based on Nested Dual-Memory Deep Deterministic Policy Gradient

Though there are extensive works on deep reinforcement learning (DRL) for robotics, sequential trajectory generation for multiprocess robotic tasks based on DRL is yet to be explored. In this article, the multiprocess task is formulated as a Markov decision process, and a nested dual-memory deep deterministic policy gradient algorithm with dynamic criteria is proposed, to generalize the traditional trajectory planning with predefined target point into a trajectory exploration problem aiming at a target area without solving inverse kinematics. First, a dual-memory architecture with local-to-global strategy is introduced to enhance the performance. Second, a novel nested architecture is proposed to generate sequential trajectory segments successively and asynchronously for the multiprocess task. Third, a compound reward system is designed and a weight coefficient matrix is adopted to balance the position control and the orientation control based on Tait–Bryan angles. In addition, a virtual twin system is established to promote the training efficiency, where the trajectory generated in simulation can be directly applied to the real physical platform. Finally, experimental results on both simulated and real-world applications have verified the performance of the proposed approach.

Principal Component Tracking for Massive MIMO Channels in High Mobility Scenarios With Diagonal Step Size Matrix

The present article proposes a novel adaptive algorithm with an optimal diagonal step size matrix for multi-dimensional channel principal component tracking in two dimensional massive multiple-input and multiple-output (M-MIMO) systems. First, we prove that the weighted subspace algorithm globally converges to the stationary stochastic process’ major eigenvectors. Then, using the maximum likelihood criterion, we optimize the weight coefficient matrix and derive the convergent condition for the step size range in order to maintain the algorithm stability. To accelerate the convergence, we initially suggest the diagonal step size matrix for multi-dimensional eigenvector tracking. Simultaneously, an optimal diagonal step size matrix is derived, which not only accelerates the convergence speed distinctively but also improves the tracking of multi-dimensional eigenvectors. Moreover, the transient behavior during the adaptation process is investigated in a straight-forward way and the relationship between the convergence time constant and the eigenvalues of the received signals is uncovered. Finally, simulations reveal that the proposed approach outperforms established algorithms such as Oja <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$^{\prime}\text{s}$ </tex-math></inline-formula> , Delmas and gradient descent algorithms. This approach establishes a sound foundation for tracking channel state information in M-MIMO systems with great mobility.

Noise Suppression of Towed Line Array Sonar Platform Based on Spatial Filtering Technique

The problem of interference caused by the radiated noise of the towed line array sonar platform to the hydrophone line array and the formation of detection blind area in the direction of the end head of the line array is addressed. In this paper, the optimal spatial matrix filter is designed using the plane wave array manifold, the platform radiation noise copy vector, and the weighted coefficient matrix. Through iteration of the weighted coefficient matrix, the null matrix filter achieves the effect of equalizing the response of each direction of the far-field plane wave signal under the condition of zero response constraint of the platform radiation noise. The method of generalized singular value simplified decomposition is used to analyze and check the correctness of the optimal solution of the filter in theory. From the simulation results, the filter with iterative calculation is more effective in processing far-field plane waves and the signal has lower distortion.

Dynamic Lane Tracking Control of the Commercial Vehicle Based on RMPC Algorithm Considering the State of Preceding Vehicle

In order to improve the adaptability of the lane keeping control system to complex environments, a dynamic lane tracking control strategy of the commercial vehicle based on the robust model predictive control (RMPC) algorithm is proposed considering the state of the preceding vehicle. An RMPC controller is designed with path deviation and control increment as the objective function. The model predictive control problem is transformed into a min–max optimization problem. The linear matrix inequality (LMI) is used for the optimal solution to obtain the optimal control quantity. The strategy to improve the safety and comfort dynamically in the process of lane keeping is designed by adjusting the weight coefficient matrix of RMPC based on fuzzy theory. The results of the simulation and HiL test show that the RMPC controller can meet the requirement of adjusting the lane tracking process dynamically according to the state of the preceding vehicle, which keeps the balance between safety and comfort.

Study on the LQR Control of High-speed Elevator Car Horizontal Vibration Based on the Jumping Inertia Weight Particle Swarm Optimization

To reduce the horizontal vibration of a high-speed elevator car caused by the excitation of a guide rail more effectively, a differential magnetic suspension active guide shoe based on the principle of the differential magnetic suspension actuator is designed. Then the dynamic model of the active vibration damping system of the elevator car is established and an LQR controller is designed to reduce the horizontal vibration of the car. To search the weighting coefficient matrix of the LQR controller more efficiently, an algorithm named Jumping inertia Weight Particle Swarm Optimization (JWPSO) algorithm is proposed, and the frequently used fitness function verifies the optimization effect of the JWPSO algorithm. The weighting coefficient matrix of the LQR controller is optimized using the JWPSO algorithm. Finally, the impact of the JWPSO algorithm-optimized LQR controller is verified by MATLAB. The simulation result shows that the designed active vibration controller can effectively attenuate the horizontal vibration of the elevator car, and the control effect is significantly better than the LQR controller optimized by the GA algorithm and the H-inf controller optimized by LMI. This paper provided a new method for horizontal vibration reduction of the high-speed elevator car.

The Improved Three-Step Semi-Empirical Radiometric Terrain Correction Approach for Supervised Classification of PolSAR Data

The radiometric terrain correction (RTC) is an essential processing step for supervised classification applications of polarimetric synthetic aperture radar (PolSAR) over mountainous areas. However, the current angular variation effect (AVE) correction methods of three-step RTC processing are difficult to apply to PolSAR supervised classification because of the problem of interdependence between AVE correction and classification. To address this issue, based on the three-step semi-empirical RTC approach, we propose an improved AVE correction method suitable for the supervised classification of PolSAR. We make full use of the prior knowledge required for supervised classification and RTC processing, that is, samples and elevation data, to calculate the parameters of AVE correction by constructing a weight coefficient matrix. GaoFen-3 QPSI (C-band, quad-polarization) data were used to verify the proposed method. Experimental results showed that the proposed method is available and effective for PolSAR supervised classification. The new method can effectively remove the AVE effect in the PolSAR image, and the overall accuracy of PolSAR supervised classification can be improved about 9% compared to that without AVE correction. For the fine classification of forest types, the AVE correction can improve the classification accuracy by about 20%.

Finite control‐set model predictive direct speed control of a PMSM drive based on the Taylor series model

In the traditional finite control-set model predictive direct speed control (FCS-MPDSC), the coupling items in the current equation have been ignored because of predictive complexity, leading to imperfect prediction accuracy and control performance. Meanwhile, the current error and speed error are both included in the weighted sum cost function. The two types of errors lead to the difficulty of weight coefficients' distribution, which can only rely on subjective experiences. Thus, an FCS-MPDSC strategy based on the 2nd-Taylor-series model with quadratic cost function is proposed. The dynamic performance has been greatly improved because the input variables of the Taylor series model are unified as the basic voltage vector with the same control period. And the static error is also reduced as the weight coefficient matrix of the quadratic cost function is solved offline. On this basis, the verification experiments are carried out. The results show that a better dynamic and static performance is achieved at the same time by the proposed strategy.

Soil salinity inversion based on differentiated fusion of satellite image and ground spectra

Due to the wide availability of remote sensing data from different sensors and platforms, soil salinity inversion based on the fusion of multisource remote sensing data is becoming a reality. However, existing fusion methods mainly use the average relationship from samples, which does not consider the differences in the relationships among samples. In this paper, a differentiated fusion method for determining satellite and ground spectral variables of soil salinity according to the differences among samples is proposed to increase the regional inversion precision. Nonnegative matrix factorization was employed to decompose soil salinity spectral variables from the Sentinel-2A Multi-Spectral Instrument (MSI) image and the simulated spectra on ground spectra. Then, the base spectra matrix of soil salinity was from the simulated data, and the weight coefficient matrix was obtained from the Sentinel-2A MSI data as the differentiated correction coefficients. By multiplying the base matrix and weight matrix, the spectral variables were reconstructed. The results indicate that this differentiated fusion method can not only enhance the correlation between soil salinity and Sentinel-2A MSI data but also improve the precision of regional soil salinity inversion models. For the differentiated fused model, the validation R2, RMSE, and RPD were 0.71, 7.02 g/kg, and 1.49, respectively; compared with the unfused model, the validation R2 increased by 0.09, the RMSE decreased by 0.80 g/kg, and the RPD increased by 0.18. Furthermore, the differentiated fused model performed better than the average-ratio adjusted model. These findings have practical implications for the use of multisource optical remote sensing data for regional soil salinity mapping and analysis.

Impulsive stabilization and stability analysis for Gilpin–Ayala competition model involved in harmful species via LMI approach and variational methods

Firstly, a dynamic analysis for reaction–diffusion Gilpin–Ayala competition model involved in harmful species is considered under Dirichlet boundary value condition. Existence of multiple stationary solutions is verified by way of Mountain Pass lemma, and the local stability result of the null solution is obtained by employing linear approximation principle. Secondly, the authors utilize variational methods and linear matrix inequality (LMI) technique to deduce the LMI-based global exponential stability criterion on the null solution which becomes the unique stationary solution of a Markovian jumping ecosystem with delayed feedback under a reasonable boundedness assumption on population densities. Particularly, LMI criterion is involved in free weight coefficient matrix, which reduces the conservatism of the algorithm. In addition, a new impulse control stabilization criterion is also derived, in which no differentiable assumptions on time-delayed functions are proposed. Finally, three numerical examples show the effectiveness of the proposed methods. It is worth mentioning that the obtained stability criteria of null solution presented some useful hints on how to eliminate pests and bacteria.

One-dimensional Block Matching Algorithm in Video Image Motion Estimation

<p indent=0mm>Motion estimation is a basic problem in the fields of video image compression and video image restoration. The traditional block matching methods have good search quality, but the search speed is not fast enough. Aiming at the shortcomings of the search speed in the traditional block matching methods, we proposed a fast one-dimensional block matching motion estimation algorithm. Firstly, the motion vector is orthogonally decomposed, and the two-dimensional matching block is reduced by a special weight coefficient matrix to obtain two sets of one-dimensional feature matrices. Then the one-dimensional three step search method is selected as the search strategy. The sum of absolute differences criterion is used as the matching criterion. The two sets of one-dimensional feature matrices are used to search for the two components of the matching motion vector. Finally the two components are formed into a complete motion vector. The results of multiple sets of comparative experiments show that the search speed of motion estimation is significantly improved while the algorithm is guaranteed to quantitatively evaluate the PSNR. The higher the video definition and the larger the pixel size of the matching block, the better the algorithm can improve the search speed of motion estimation.

Big Data Interprets US Opioid Crisis

Since 2010, there has been a new round of drug crises in the United States. The abuse of opioids has led to a sharp increase in the number of people involved in drug crimes in the United States. There is an urgent need to explore solutions to the drug crisis in the United States. In this paper, the model of in-depth analysis is established under the condition of obtaining the opioid data and the influence factor data of the large sample of five state[1]. In the first part, we use the Highway Safety Research Institute model based on the differential equation model to predict the initial value, find the initial position of the drug transfer, and obtain the curve of the number of different groups over time by fitting the data, so that the curves can be predicted the changing trends of the groups in the future. It was found that in Kentucky State, the county’s most likely to start using opioids were Pike and Bale. In Ohio, the county’s most likely to start using opioids are Jackson and Scioto. In Pennsylvania State, Mercer and Lackawanna are the counties most likely to start using opioids. Martinsville and Galax are the counties where Virginia State is most likely to start using opioids. Logan and Mingo are the counties where West Virginia State is most likely to start using opioids. In the second part, the gray prediction model is used to further analyze the time series of each factor, the maximum likelihood estimation method is used to obtain the weight of each factor, and the weight coefficient matrix is used to simulate the multivariate regression equation, and the factors that have the greatest influence on opioid abuse are educational background and family composition. In the third part, the hypothesis test model of two groups (the data type is proportional) is used to verify the difference between the influence factors (including the predicted values) in the first two parts of the states, thus verifying the feasibility between them. At the same time, we put forward a few suggestions to combine the current situation in the United States with the CDC data. We believe that in order to address the opium crisis, the U.S. government needs to strengthen not only oversight of doctors' prescriptions, but also make joint efforts of all sectors of society to fundamentally reduce the barriers to the use of opioids.

Vulnerability evaluation of high-rise earthquake-proof buildings in high-pressure area

In order to solve this problem, the vulnerability evaluation method of high-rise seismic buildings in high-pressure area based on fuzzy comprehensive evaluation is proposed. According to the knowledge of earthquake engineering, the relationship between magnitude and intensity and the law of intensity attenuation are analysed. Based on the seismic risk and the vulnerability of the isolation layer of the building structure, the failure evaluation criteria are established. The evaluation indexes of structural vulnerability are selected, and the weight coefficients of each index are determined by using tomographic method, and the corresponding matrix is obtained. The fuzzy comprehensive evaluation method is used to evaluate each index. According to the fuzzy relation matrix and index weight coefficient matrix of evaluation index and vulnerability level, the vulnerability of building structure is calculated and evaluated. Experimental results show that the evaluation performance of this method is better than that of the experimental comparison method.

Weighted Decoding for the Competence Reliability Problem of ECOC Multiclass Classification.

Error-Correcting Output Codes has become a well-known, established technique for multiclass classification due to its simplicity and efficiency. Each binary split contains different original classes. A noncompetent classifier emerges when it classifies an instance whose real class does not belong to the metasubclasses which is used to learn the classifier. How to reduce the error caused by the noncompetent classifiers under diversity big enough is urgent for ECOC classification. The weighted decoding strategy can be used to reduce the error caused by the noncompetence contradiction through relearning the weight coefficient matrix. To this end, a new weighted decoding strategy taking the classifier competence reliability into consideration is presented in this paper, which is suitable for any coding matrix. Support Vector Data Description is applied to compute the distance from an instance to the metasubclasses. The distance reflects the competence reliability and is fused as the weight in the base classifier combination. In so doing, the effect of the competent classifiers on classification is reinforced, while the bias induced by the noncompetent ones is decreased. Reflecting the competence reliability, the weights of classifiers for each instance change dynamically, which accords with the classification practice. The statistical simulations based on benchmark datasets indicate that our proposed algorithm outperforms other methods and provides new thought for solving the noncompetence problem.