While existing flow-based models have the potential for data enhancement and intelligent fault diagnosis of bearings, these methods are mainly based on the time domain to learn features. In general, time-domain and fr...
详细信息
ISBN:
(数字)9798350360585
ISBN:
(纸本)9798350360592
While existing flow-based models have the potential for data enhancement and intelligent fault diagnosis of bearings, these methods are mainly based on the time domain to learn features. In general, time-domain and frequency-domain features are important for recognizing faults. Especially, the spectral features in the frequency domain can reveal highly distinct patterns associated with fault types. Therefore, learning spectral feature helps to obtain comprehensive fault information and achieve high fault diagnosis accuracy. In this work, a spectral flow model framework is developed to learn and generate realistic data from mechanical sensor signals. The performance of fault diagnosis under unbalanced data is improved by learning the information contained in the spectral features, which are utilized to generate insufficient fault data. Experimental results show that the present method can generate a large number of samples with higher similarity and better diversity on the bearing dataset. The generated samples may be used to expand the limited dataset and significantly improve the accuracy of the bearing fault diagnosis task.
We consider the event-triggered state estimation of a finite-state hidden Markov model with a general stochastic event-triggering condition. Utilizing the change of probability measure approach and the event-triggered...
详细信息
ISBN:
(纸本)9781479978878
We consider the event-triggered state estimation of a finite-state hidden Markov model with a general stochastic event-triggering condition. Utilizing the change of probability measure approach and the event-triggered measurement information available to the estimator, analytical expressions for the conditional probability distributions of the states are obtained, based on which the minimum mean square error event-based state estimates are further calculated. We show that the results also cover the case of packet dropout, under a special parameterization of the event-triggering conditions. With the results on state estimation, a closed-form expression of the average sensor-to-estimator communication rate is also presented. The effectiveness of the proposed results is illustrated by a numerical example and comparative simulations.
An adaptive fuzzy observer for nonlinear magnetic levitation system with uncertain friction coefficient is proposed. First, a T-S fuzzy model of the nonlinear magnetic levitation system is proposed. An observer is giv...
详细信息
An adaptive fuzzy observer for nonlinear magnetic levitation system with uncertain friction coefficient is proposed. First, a T-S fuzzy model of the nonlinear magnetic levitation system is proposed. An observer is given by LMI method and the Lyapunov method. Through adding an auxiliary variable, it relaxes the constraint of the design of observer. The proposed approach is a more relaxed condition than others. Simulation results show the effectiveness of this approach. The results show that the position of magnetic levitation system is estimated effectively with unknown friction coefficient. There is some reference value for a kind of nonlinear systems with unknown parameter.
A new method is proposed, through combining the algorithm of orthogonal discriminant linear local tangent space alignment (ODLLTSA) and the support vector machine (SVM), to improve the accuracy of recognizing door pla...
详细信息
A new method is proposed, through combining the algorithm of orthogonal discriminant linear local tangent space alignment (ODLLTSA) and the support vector machine (SVM), to improve the accuracy of recognizing door plate numbers. The feature of door plate characters is first extracted by the ODLLTSA and then this extracted feature is used to train the SVM classifier. Finally, the new plate characters are classified by the trained SVM. Using the algorithm, a high recognition rate can be achieved. Experimental results show that this method is effective and robust in the real applications.
We address robust stabilization problem for networked controlsystems with nonlinear uncertainties and packet losses by modelling such systems as a class of uncertain switched systems. Based on theories on switched Ly...
详细信息
ISBN:
(纸本)9781424445233
We address robust stabilization problem for networked controlsystems with nonlinear uncertainties and packet losses by modelling such systems as a class of uncertain switched systems. Based on theories on switched Lyapunov functions, we derive the robustly stabilizing conditions for state feedback stabilization and design packet-loss dependent controllers by solving some matrix inequalities. A numerical example and some simulations are worked out to demonstrate the effectiveness of the proposed design method.
In order to improve real-time performance of the fire controlsystem, a ballistic resolving method based on the improved particle swarm optimization (PSO) algorithm is proposed, which improves the response speed of th...
详细信息
In order to improve real-time performance of the fire controlsystem, a ballistic resolving method based on the improved particle swarm optimization (PSO) algorithm is proposed, which improves the response speed of the system and provides a convenient extension to parallel computing on multicore platforms. First, particles are generated and initialized around the pre-estimated aiming angle. Then each particle is evaluated by an objective function composed of the ballistic differential equation etc. Finally, the position and velocity of particle swarm are updated. In order to accelerate the convergence speed, the correction angle of the global best particle obtained by Zhou's iterative and correction formula is used to guide the update of particle swarm. Experimental results show that the calculation speed is twice that of the iterative and correction method, and the convergence speed of particle swarm is 1.5 times that of the conventional PSO algorithm. Moreover, the proposed method is fully compatible with parallel computing and can further shorten execution time on multicore platforms.
The fault diagnosis for a class of widely used digital parallel output optical encoder were focused. After definition of the optical encoder, the main features of the optical encoder's output data were analyzed. A...
详细信息
The fault diagnosis for a class of widely used digital parallel output optical encoder were focused. After definition of the optical encoder, the main features of the optical encoder's output data were analyzed. A fault diagnosis method which did not rely on the system model where optical encoder used was proposed. The changes of optical encoder's output data were analyzed. Then, the inherent characteristics were calculated. The fuzzy logic was utilized to determine the fault type and locate the fault location. Theoretical analysis and experimental results show that this method can diagnose and isolate optical encoder fault accurately without disassembly.
This brief addresses the time-varying formation control problem with collision avoidance for second-order multi-agent systems. By taking both distances and velocities between agents into account, a novel collision avo...
详细信息
In order to describe precisely the dynamics and friction nonlinearity of servo systems, a novel direct identification method for nonlinear continuous model is proposed. The sampled input-output data and logic data cor...
详细信息
In order to describe precisely the dynamics and friction nonlinearity of servo systems, a novel direct identification method for nonlinear continuous model is proposed. The sampled input-output data and logic data corresponding to the velocity direction are chosen as the identification data. Through equivalent transformation, the unknown parameters are removed to the linear part of the model. Then the identification method based on state variable filter is utilized to determine the unknown parameters. Subsequently, the nonlinear continuous model of the servo system is obtained. The effectiveness of the proposed method is demonstrated by simulations and identification experiments on the two axis servo table. Both the simulations and experimental results show that, with the proposed method, the accurate nonlinear continuous model can be obtained even under sensor noises, which gives an accurate description of the system dynamics.
In this paper an Ad hoc networks are established for the communication among unmanned air vehicle (UAV) group corporation. It is brought out that reference point group mobility model (RPGM) could simulate the behavior...
详细信息
In this paper an Ad hoc networks are established for the communication among unmanned air vehicle (UAV) group corporation. It is brought out that reference point group mobility model (RPGM) could simulate the behavior of UAV group to analyze the efficiency of router algorithms as to the movement of UAV node. The result shows that the Ad hoc networks can solve the communication problem of UAV group, and also find out that the ADOV algorithm has data large throughout and more adaptive to the variation of topology, the DSR has a lighter routing load. The efficiency of router algorithm also changes as to the node speed. As a result the routing algorithm should be chosen according to the different tactical mission of UAV group, and the designs of corporation method should consider the changing of network efficiency as to node moving speed.
暂无评论