An adaptive structure radial basis function (RBF) network model is proposed in this paper to model nonlinear processes with operating point migration. The recursive orthogonal least squares algorithm (rols) is adopted...
详细信息
An adaptive structure radial basis function (RBF) network model is proposed in this paper to model nonlinear processes with operating point migration. The recursive orthogonal least squares algorithm (rols) is adopted to select new centers on-line, as well as to train the network weights. Based on the R matrix in the orthogonal decomposition, an initial center bank is formed and updated in each sample period. A new learning strategy is proposed to gain information from the new data for network structure adaptation. A center grouping algorithm is also developed to divide the centers into active and non-active groups, so that a structure with a smaller size is maintained in the final network model used for output prediction. The proposed REF model is evaluated and compared with the three existing adaptive structure RBF networks by modeling a nonlinear time-varying numerical example. Simulation results demonstrate that the proposed algorithm has several advantages in term of the adaptive tracking ability and a better recovery speed over the existing methods during the migration of system's operating point. (C) 2015 Elsevier B.V. All rights reserved.
In this paper, a fuzzy wavelet network is proposed to approximate arbitrary nonlinear functions based on the theory of multiresolution analysis (MRA) of wavelet transform and fuzzy concepts. The presented network comb...
详细信息
In this paper, a fuzzy wavelet network is proposed to approximate arbitrary nonlinear functions based on the theory of multiresolution analysis (MRA) of wavelet transform and fuzzy concepts. The presented network combines TSK fuzzy models with wavelet transform and rols learning algorithm while still preserve the property of linearity in parameters. In order to reduce the number of fuzzy rules, fuzzy clustering is invoked. In the clustering algorithm, those wavelets that are closer to each other in the sense of the Euclidean norm are placed in a group and are used in the consequent part of a fuzzy rule. Antecedent parts of the rules are Gaussian membership functions. Determination of the deviation parameter is performed with the help of gold partition method. Here, mean of each function is derived by averaging center of all wavelets that are related to that particular rule. The overall developed fuzzy wavelet network is called fuzzy wave-net and simulation results show superior performance over previous networks. The present work is complemented by a second part which focuses on the control aspects and to be published in this journal([17]). This paper proposes an observer based self-structuring robust adaptive fuzzy wave-net (FWN) controller for a class of nonlinear uncertain multi-input multi-output systems. (C) 2010 Elsevier B.V. All rights reserved.
This paper proposes the architecture of Radial Basis Function (RBF) neural networks for the recognition of different digital modulated signals. Recursive Orthogonal Least Squares (rols) algorithm is used not only for ...
详细信息
ISBN:
(纸本)0780378652
This paper proposes the architecture of Radial Basis Function (RBF) neural networks for the recognition of different digital modulated signals. Recursive Orthogonal Least Squares (rols) algorithm is used not only for calculating the weights of the network, but also for choosing RBF neural networks centers sequentially after network training, according to minimizing the output error. The final network models can achieve acceptable accuracy with significant reduction in the number of required, centers without retraining. The trained networks have the ability to recognize most digital modulated signals, such as 2ASK 4ASK, 2FSK 4FSK, 2PSK, 4PSK and other hybrid multiplex modulated signals. The simulation results demonstrate the validity and practicability of this approach.
For time-varying processes a fixed neural network (NN) model trained off-line cannot be used for on-line fault detection, as the NN cannot model the process dynamics change and therefore the modelling error will incre...
详细信息
For time-varying processes a fixed neural network (NN) model trained off-line cannot be used for on-line fault detection, as the NN cannot model the process dynamics change and therefore the modelling error will increase. When sensor faults are detected, a parallel model should be used for multi-step ahead prediction. In this case the time-varying effects will result in modelling error diverging. An on-line fault detection strategy using adaptive parallel radial basis function network (RBFN) is proposed in this paper. The RBFN is trained on-line using the recursive orthogonal least squares (rols) algorithm until a fault is detected. An error index is formulated to distinguish the dynamics change from the faults. The method is applied to a three-input three-output chemical process to detect the simulated faults on the real data. The results indicate the feasibility of the method and the applicability to the real dynamic processes.
An adaptive radial basis function (RBF) neural network model is developed in this paper for nonlinear systems using the recursive orthogonal least squares (rols) algorithm. The model is used in a nonlinear model predi...
详细信息
An adaptive radial basis function (RBF) neural network model is developed in this paper for nonlinear systems using the recursive orthogonal least squares (rols) algorithm. The model is used in a nonlinear model predictive control (NMPC). The developed adaptive NMPC is applied to a chemical reactor rig. On-line control performance is presented and it demonstrates superiority over the fixed parameter PID control.
Greenhouse environment models easily fitted strong noise data,and its' generalization *** this paper,rols(Regularized Orthogonal Least Squares) algorithm effectively decreased the influence of noise data,and autom...
详细信息
Greenhouse environment models easily fitted strong noise data,and its' generalization *** this paper,rols(Regularized Orthogonal Least Squares) algorithm effectively decreased the influence of noise data,and automatically designed smaller NN ***(Particle Swarm Optimization) algorithm optimized the parameters of *** was experimented with spring environment data of northern greenhouse in *** results show:compared with model based on OLS algorithm,this model is of smaller NN structure,mean error of temperature and humidity respectively decreases 0.0008 ℃ and 0.0004%RH,this model is better on approximation and *** model is beneficial to design control scheme and structure of the northern greenhouse.
暂无评论