A design of nonlinear-dynamic observer is proposed for determining the states of a nonlinear system. The design method uses a multi-layered feed-forward neural network (MFNN) to approximate the nonlinear Kalman gain. ...
详细信息
A design of nonlinear-dynamic observer is proposed for determining the states of a nonlinear system. The design method uses a multi-layered feed-forward neural network (MFNN) to approximate the nonlinear Kalman gain. Two different criteria are proposed for the network training. The training is based on a gradient descent algorithm that uses block partial derivatives. Simulation results on Van der Pol's equation and the classical inverted pendulum model are presented to validate the usefulness of the scheme.
The conventional back-propagation algorithm is basically a gradient-descent method, it has the problems of local minima and slow convergence. A new generalized back-propagation algorithm which can effectively speed up...
详细信息
The conventional back-propagation algorithm is basically a gradient-descent method, it has the problems of local minima and slow convergence. A new generalized back-propagation algorithm which can effectively speed up the convergence rate and reduce the chance of being trapped in local minima is introduced. The new back-propagation algorithm is to change the derivative of the activation function so as to magnify the backward propagated error signal, thus the convergence rate can be accelerated and the local minimum can be escaped. In this letter, we also investigate the convergence of the generalized back-propagation algorithm with constant learning rate. The weight sequences in generalized back-propagation algorithm can be approximated by a certain ordinary differential equation (ODE). When the learning rate tends to zero, the interpolated weight sequences of generalized back-propagation converge weakly to the solution of associated ODE.
A new fuzzy on-line identification algorithm for a single input/single output continuous-time nonlinear dynamic system is presented. This method combines the conventional on-line identification with fuzzy logic system...
详细信息
A new fuzzy on-line identification algorithm for a single input/single output continuous-time nonlinear dynamic system is presented. This method combines the conventional on-line identification with fuzzy logic system. The nonlinear system is approximated by a set of fuzzy rules that describe the local linear dynamic in each subspace formed by fuzzifying the input and output space. The continuous-time fuzzy input-output model is identified on-line by using the input and output measurements. A fuzzy identification algorithm has been developed and a convergence analysis is carried out. Simulation studies have demonstrated that this fuzzy on-line identifier can match the time-varying nonlinear system within +/-5% accuracy. (C) 1999 Elsevier Science B.V. All rights reserved.
Artificial neural networks are useful tools for pattern recognition because they realize nonlinear mapping between input and output spaces. This ability is tuned by supervised learning methods such as back-propagation...
详细信息
Artificial neural networks are useful tools for pattern recognition because they realize nonlinear mapping between input and output spaces. This ability is tuned by supervised learning methods such as back-propagation. In the supervised learning methods, desired outputs of the neural network are needed. However, the desired outputs are usually unknown in unpredictable environments. To solve this problem, this paper presents a self-supervised learning system for category detection. This system learns categories of objects automatically by integrating information from several sensors. We assume that these sensory inputs are always ambiguous patterns that include some noises according to deformations of the objects. After the learning, the system recognizes objects, also controlling the priority of each sensor, according to the deformation of the sensory input pattern. In the simulation, the system is applied to several learning and recognition tasks using artificial or actual sensory inputs. In all tasks, the system found the categories. Particularly, we applied the new system to the learning of five Japanese vowels with the corresponding shapes of the mouth. As result, the system became to yield specific outputs corresponding to each vowel. (C) 1999 Elsevier Science Ltd. All rights reserved.
This paper presents an automatic current-control method for magnet cranes for thick steel plate yard automation. In moving the steel plates from stack to stack or from stack to shipping truck, it is difficult to lift ...
详细信息
This paper presents an automatic current-control method for magnet cranes for thick steel plate yard automation. In moving the steel plates from stack to stack or from stack to shipping truck, it is difficult to lift the correct number of steel plates because the dimensions of the stacked steel plates are in general different, and their other parameters are not completely known and may even vary in a nonlinear fashion. In this paper, recursive form flux equations are first derived for the thick steel plates, and a current equation is then determined for the magnet coil. Based on these equations, an adaptive current predictor is developed to exert the right amount of current on the magnet, so as to lift the correct number of steel plates. When the initial trial with this current fails, a current tuning method is introduced to adjust the current and lift the correct number of plates. The developed magnet current controller has been successfully tested on data obtained from the storage yard at the Pohang Iron & Steel Co. (POSCO). (C) 1998 Published by Elsevier Science Lid. All rights reserved.
In this letter, we investigate the interactions of front-end feature extraction and back-end classification techniques in nonstationary state hidden Markov model (NSHMM) based speech recognition. The proposed model ai...
详细信息
In this letter, we investigate the interactions of front-end feature extraction and back-end classification techniques in nonstationary state hidden Markov model (NSHMM) based speech recognition. The proposed model aims at finding an optimal linear transformation on the mel-warped discrete Fourier tranform (DFT) features according to the minimum classification error (MCE) criterion. This linear transformation, along with the NSHMM parameters, are automatically trained using the gradientdescent method. An error rate reduction of 8% is obtained on a standard 39-class TIMIT phone classification task in comparison with the MCE-trained NSHMM using conventional preprocessing techniques.
This paper presents an intelligent PID controller based on a gradientdescent learning algorithm. A BP neural network is needed to learn the characteristics of the dynamic systems. The possibility of using neural netw...
详细信息
ISBN:
(纸本)0819420123
This paper presents an intelligent PID controller based on a gradientdescent learning algorithm. A BP neural network is needed to learn the characteristics of the dynamic systems. The possibility of using neural network models directly within a model-based predictive control strategy is also considered by making use of an on-line optimization routine to determine the future inputs that will minimize the deviation between the desired and predicted outputs. The controller's structure and the learning algorithm are very simple and easily realized. It can also replace the traditional PID controller, control the complex systems, and require neither process model nor more tuning parameters.
暂无评论