作者:
Vukic, Z.Velagic, J.University of Zagreb
Faculty of Electrical Engineering and Computing Department of Control and Computer Engineering in Automation Unska 3 ZagrebHR-10000 Croatia University of Sarajevo
Faculty of Electrical Engineering Skenderija 70 SarajevoBH-71000 Bosnia and Herzegovina
In this paper a comparative analysis of Mamdani and Sugeno type fuzzy autopilots for ships is given. Both autopilots have two inputs: the heading signal and the yaw rate signal, and only one output: command rudder ang...
详细信息
Contemporary trends show a marked tendency for solution of problems of dynamic analysis of drive systems as early as the design stage. 11 shows that during the modelling of complex drive systems it is neccssaiy to res...
详细信息
Tuning procedures for activation functions significantly increases the flexibility and the nonlinear approximation capability of feedforward neural networks in supervised learning tasks. As a consequence, the learning...
详细信息
Tuning procedures for activation functions significantly increases the flexibility and the nonlinear approximation capability of feedforward neural networks in supervised learning tasks. As a consequence, the learning process presents a better performance, with the final state of the neural network being kept away from undesired saturation regions. Based on a hybrid architecture combining a gradient strategy with a fuzzy decision model, an auto-tuning algorithm is derived to adjust additional parameters associated with the activation functions. The other conventional parameters, the connection weights between layers, are adjusted using a powerful second-order approach based on a conjugate gradient algorithm. To demonstrate the performance of the proposed method we compare this technique with the standard algorithm and with an auto-tuning strategy based solely on the gradient descent method. The three algorithm are applied to several artificial and real world benchmarks.
Friction appears in the bearings and reduction gear of controlled electrical drives and affects the quality of the position, speed or force control. The servosystem performance can be improved by implementation of acc...
详细信息
Friction appears in the bearings and reduction gear of controlled electrical drives and affects the quality of the position, speed or force control. The servosystem performance can be improved by implementation of acceleration feedback. For the reason that acceleration measurement is often either not possible or practical, the implementation of acceleration estimation by an observer or differentiation of speed is proposed. Another approach to friction influence compensation is based on the disturbance observer. Its advantage is that it can be applied generally to the estimation of various disturbances. This paper deals with these friction compensation methods which are not based on friction model, but on the compensation of friction as a disturbance. The efficiency of these methods in a servodrive with Stribeck friction is compared by computer simulation and experimentally on a servosystem laboratory model.
This paper describes B-spline interpolation and compares it with other reconstruction methods, especially in three-dimensional space. We first consider the B-spline bases in the terms of convolution in signal processi...
详细信息
This paper describes B-spline interpolation and compares it with other reconstruction methods, especially in three-dimensional space. We first consider the B-spline bases in the terms of convolution in signal processing. The presented analysis requires careful usage of continuous and discrete representation of B-spline. Emphasis is given to the important difference between B-spline interpolation and approximation. The difference is shown through frequency domain analysis, so we derive frequency responses of the B-spline interpolation and approximation. We conclude by demonstrating the use of several reconstruction filters and appropriate gradient estimators in volume rendering. Exact reconstruction in volume visualization is very important in many industrial applications, such as material cavity control.
Application of genetic algorithm and genetic programming to identification of a nonlinear vibration system is presented. Both the theoretical groundwork and experimental results are included. The genetic algorithm is ...
详细信息
The volume data is generally in the form of the large array of numbers. In order to render the object hidden in the volumetric data, we need to reconstruct or interpolate data values between the samples. The novelty p...
详细信息
The volume data is generally in the form of the large array of numbers. In order to render the object hidden in the volumetric data, we need to reconstruct or interpolate data values between the samples. The novelty presented in this paper is B-spline interpolation in the volumetric space. We show that this approach is better than currently used methods. We also present a hybrid approach, analyze this approach in frequency domain and compare it to B-spline interpolation. To enhance the quality during the volume visualization process it is important to enhance the quality of the reconstruction. It is of crucial importance to explore different undesired effects. If better reconstruction is performed the more accurate result of volume visualization process is achieved.
Presents the latest results of sensing the cutting process on the basis of acoustic emission (AE) signals and some particularities in the further development of a monitoring model for the finish turning process. Due t...
详细信息
Presents the latest results of sensing the cutting process on the basis of acoustic emission (AE) signals and some particularities in the further development of a monitoring model for the finish turning process. Due to nonlinearity, the large number of influencing parameters and missing information in AE data, artificial neural networks were chosen as a monitoring decision tool. The problem of accuracy in predicting the surface roughness on the basis of AE-because of the mutual interdependence of the data-requires a special procedure for building a neural network model. The final aim of such an approach is presented as improvements in learning or considerable reduction in error prediction. Further development of the monitoring model has the goal of building a so-called intelligent sensor, which should be able to perform the signal conditioning and feature extraction process.
In this paper the robust fault detection and isolation (FDI) problem in linear systems, which are subjected to multiple faults and time varying system perturbations is discussed. The detection and isolation problem is...
详细信息
In this paper the robust fault detection and isolation (FDI) problem in linear systems, which are subjected to multiple faults and time varying system perturbations is discussed. The detection and isolation problem is treated by using algebraic methods and geometric design concepts. The idea of the classical Luenberger observer is generalized in an attempt that residual generators for solving the robust FDI problem can be designed.
In this paper the terminal sliding mode technique is used for solving the finite time output tracking problem for non-minimum phase nonlinear systems. A novel idea of stabilizing zero dynamics of the non-minimum phase...
详细信息
In this paper the terminal sliding mode technique is used for solving the finite time output tracking problem for non-minimum phase nonlinear systems. A novel idea of stabilizing zero dynamics of the non-minimum phase system is proposed. The terminal sliding mode control is employed to realize the finite time convergence.
暂无评论