In this paper, the authors report on the design, simulation and validation of an adaptive neuro-fuzzy inference system (ANFIS) based power system stabilizer (PSS) for a single-machine-infinite-bus (SMIB) and a multi-m...
详细信息
ISBN:
(纸本)9781861353429
In this paper, the authors report on the design, simulation and validation of an adaptive neuro-fuzzy inference system (ANFIS) based power system stabilizer (PSS) for a single-machine-infinite-bus (SMIB) and a multi-machine power system and investigate its performance in damping low frequency local and inter-area oscillations. The design employs a first order Sugeno fuzzy model, whose parameters are tuned off-line through hybrid learning algorithm. This algorithm is a combination of least square estimator and error backpropagation method. The performance of the ANFIS-based PSS is observed through digital simulation for both SMIB and multi-machine systems. Finally the results are compared with conventional fuzzy PSS performances. It is observed that ANFIS-based PSS is playing more satisfactory role in damping local and inter-area oscillations, which proves its effectiveness in small-signal stability.
In this paper a neural network model-based predictive control strategy for aircraft systems with unknown parameters is presented. The objective of the paper is to stabilize unknown systems by the adaptive control law ...
详细信息
In this paper a neural network model-based predictive control strategy for aircraft systems with unknown parameters is presented. The objective of the paper is to stabilize unknown systems by the adaptive control law through an optimization procedure in which a cost function representing the deviation error between set-points and the predicted outputs obtained from a neural network is minimized. Due to its capability of characterizing dynamic functional relationships and its feedback processing structure, a recurrent neural network is employed as an adaptive estimator for future state values. The neural network training is performed by the dynamic sequential recursive backpropagation learning algorithm, which allows the neural network to be trained online. It is shown that the proposed neural network learning algorithm has potential for designing flight control systems which can compensate for unpredictable changes in an aircraft dynamics over a wide range of flight conditions and other uncertainties.
This paper proposes an SAC using neural networks with offset error reduction for MIMO nonlinear systems. In this proposed method, the control input for the nonlinear plant is given by the sum of the output of a simple...
详细信息
This paper proposes an SAC using neural networks with offset error reduction for MIMO nonlinear systems. In this proposed method, the control input for the nonlinear plant is given by the sum of the output of a simple adaptive controller and the output of neural networks. The role of neural networks is to compensate for constructing a linearized model so as to minimize the output error caused by nonlinearities in the control system. The neural networks use the backpropagation algorithm for the learning process. The role of simple adaptive controller is to perform the model matching for the linear system with unknown structures to a given linear reference model. In this method, only part of the control input is fed to the PFC. Thus, the proposed method will reduce the offset error, and both of the augmented plant output and the real plant output can follow significantly close to the output of the reference model. Finally, the effectiveness of this method is confirmed through computer simulations.
In this paper, a particle swarm optimization (PSO) based camera calibration approach is presented to determine the external and internal calibration parameters from the knowledge of a given set of points in object spa...
详细信息
In this paper, a particle swarm optimization (PSO) based camera calibration approach is presented to determine the external and internal calibration parameters from the knowledge of a given set of points in object space. First, the image formation model for a pinhole camera is formulated in terms of a feed-forward neural network (NN) and then this neural network is trained using particle swarm optimization. The effect of noise and number of control points are studied in the estimation of calibration parameters. Results from our extensive study are presented to demonstrate the excellent performance of the proposed technique in terms of convergence, accuracy, and robustness.
In this work we tackle the road sign problem with reservoir computing (RC) networks. The T-maze task (a particular form of the road sign problem) consists of a robot in a T-shaped environment that must reach the corre...
详细信息
In this work we tackle the road sign problem with reservoir computing (RC) networks. The T-maze task (a particular form of the road sign problem) consists of a robot in a T-shaped environment that must reach the correct goal (left or right arm of the T-maze) depending on a previously received input sign. It is a control task in which the delay period between the sign received and the required response (e.g., turn right or left) is a crucial factor. Delayed response tasks like this one form a temporal problem that can be handled very well by RC networks. Reservoir computing is a biologically plausible technique which overcomes the problems of previous algorithms such as backpropagation through time - which exhibits slow (or non-) convergence on training. RC is a new concept that includes a fast and efficient training algorithm. We show that this simple approach can solve the T-maze task efficiently.
The quality detection of the cold-strip steel using artificial neural networks is studied. A simple Back-propagation (BP) algorithm based on error function was presented. It deals with the saturation areas that play a...
详细信息
The quality detection of the cold-strip steel using artificial neural networks is studied. A simple Back-propagation (BP) algorithm based on error function was presented. It deals with the saturation areas that play a significant role in the slow convergence of standard BP algorithm. A modified error function was constructed to make the weight adjustment to avoid falling into the saturation areas. The simulation and experiment results show the effect of improved BP algorithm on the classification of the surface defects of steel strip.
A Jordan-style cascade-correlation architecture is developed for radar signal pulse detection. The cascade-correlation learning architecture is modified to facilitate hardware implementation of the network. The networ...
详细信息
A Jordan-style cascade-correlation architecture is developed for radar signal pulse detection. The cascade-correlation learning architecture is modified to facilitate hardware implementation of the network. The network is constructed using only two hidden layers, with nodes added to the layers in a lateral fashion. Comparisons to networks trained using backpropagation and genetic algorithms indicates that the cascade-correlation architecture trains approximately 50 times faster and produces much better generalization.< >
暂无评论