A special hierarchical fuzzy neural-networks based reinforcement learning for multi-variables system

Proposes a reinforcement learning scheme based on a special Hierarchical Fuzzy Neural-Networks (HFNN)for solving complicated learning tasks in a continuous multi-variables environment. The output of the previous layer in the HFNN is no longer used as if-part of the next layer, but used only in then-part. Thus it can deal with the difficulty when the output of the previous layer is meaningless or its meaning is uncertain. The proposed HFNN has a minimal number of fuzzy rules and can successfully solve the problem of rules combination explosion and decrease the quantity of computation and memory requirement. In the learning process, two HFNN with the same structure perform fuzzy action composition and evaluation function approximation simultaneously where the parameters of neural-networks are tuned and updated on line by using gradient descent algorithm. The reinforcement learning method is proved to be correct and feasible by simulation of a double inverted pendulum system.

Tracking the events in the coverage of wireless sensor networks based on artificial neural-networks algorithms

Sensor deployment is an important problem in mobile wireless sensor networks.This paper presents a dis-tributed self-spreading deployment algorithm(SOMDA)for mobile sensors based on artificial neural-networks self-organizing maps algorithm.During the deployment,the nodes compete to track the event and cooperate to form an ordered topology.After going through the algorithm,the statistical distribution of the nodes approaches that of the events in the interest area.The performance of the algo-rithm is evaluated by the covered percentage of re-gion/events,the detecting ability and the energy equaliza-tion of the networks.The simulation results indicate that SOMDA outperforms uniform and random deployment with lossless coverage,enhancive detecting ability and signifi-cant energy equalization.

Neural-Network-Based Adaptive Finite-Time Control for a Two-Degree-of-Freedom Helicopter System With an Event-Triggering Mechanism

Helicopter systems present numerous benefits over fixed-wing aircraft in several fields of application.Developing control schemes for improving the tracking accuracy of such systems is crucial.This paper proposes a neural-network(NN)-based adaptive finite-time control for a two-degree-of-freedom helicopter system.In particular,a radial basis function NN is adopted to solve uncertainty in the helicopter system.Furthermore,an event-triggering mechanism(ETM)with a switching threshold is proposed to alleviate the communication burden on the system.By proposing an adaptive parameter,a bounded estimation,and a smooth function approach,the effect of network measurement errors is effectively compensated for while simultaneously avoiding the Zeno phenomenon.Additionally,the developed adaptive finite-time control technique based on an NN guarantees finitetime convergence of the tracking error,thus enhancing the control accuracy of the system.In addition,the Lyapunov direct method demonstrates that the closed-loop system is semiglobally finite-time stable.Finally,simulation and experimental results show the effectiveness of the control strategy.

Amplified Signal Response by Neuronal Diversity on Complex Networks

The effect of diversity on dynamics of coupled FitzHugh-Nagumo neurons on complex networks is numerically investigated, where each neuron is subjected to an external subthreshold signal. With the diversity the network is a mixture of excitable and oscillatory neurons, and the diversity is determined by the variance of the system＇s parameter. The complex network is constructed by randomly adding long-range connections （shortcuts） on a nearest-neighbouring coupled one-dimensional chain. Numerical results show that external signals are maximally magnified at an intermediate value of the diversity, as in the case of well-known stochastic resonance, burthermore, the effects of the number of shortcuts and coupled strength on the diversity-induced phenomena are also discussed. These findings exhibit that the diversity may play a constructive role in response to external signal, and highlight the importance of the diversity on such complex networks.

Global Synchronization of General Delayed Dynamical Networks

Global synchronization of general delayed dynamical networks with linear coupling are investigated. A sufficient condition for the global synchronization is obtained by using the linear matrix inequality and introducing a reference state. This condition is simply given based on the maximum nonzero eigenvalue of the network coupling matrix. Moreover, we show how to construct the coupling matrix to guarantee global synchronization of network, which is very convenient to use. A two-dimension system with delay as a dynamical node in network with global coupling is finally presented to verify the theoretical results of the proposed global synchronization scheme.

Synchronizing Complex Networks by an Adaptive Adjustment Mechanism

We propose an adaptive adjustment mechanism for synchronizing complex networks, in particular for sociological or/and biological systems. We do not take it for granted that a dynamical system is put on isolated nodes and they are coupled with each other by one （or more） variable（s）, as employed in most previous models. As a replacement, we suppose that each node may have any finite number of possible states, and their evolutions with time are determined by their nearest-neighbouring （or even second-nearest-neighbouring, etc） nodes in an adaptive adjustment mechanism. It is found that synchronization can be achieved for almost all connected networks and that the scale-free property can evidently improve the synchronizing speed.

Modelling and control PEMFC using fuzzy neural networks

Proton exchange membrane generation technology is highly efficient, clean and considered as the most hopeful “green” power technology. The operating principles of proton exchange membrane fuel cell (PEMFC) system involve thermo-dynamics, electrochemistry, hydrodynamics and mass transfer theory, which comprise a complex nonlinear system, for which it is difficult to establish a mathematical model and control online. This paper first simply analyzes the characters of the PEMFC; and then uses the approach and self-study ability of artificial neural networks to build the model of the nonlinear system, and uses the adaptive neural-networks fuzzy infer system (ANFIS) to build the temperature model of PEMFC which is used as the reference model of the control system, and adjusts the model parameters to control it online. The model and control are implemented in SIMULINK environment. Simulation results showed that the test data and model agreed well, so it will be very useful for optimal and real-time control of PEMFC system.

Design and Realization of CPW Circuits Using EC-ANN Models for CPW Discontinuities

Novel accurate and efficient equivalent circuit trained artificial neural-network （EC-ANN） models,which inherit and improve upon EC model and EM-ANN models＇ advantages,are developed for coplanar waveguide （CPW） discontinuities. Modeled discontinuities include ： CPW step, interdigital capacitor, symmetric cross junction, and spiral inductor, for which validation tests are performed. These models allow for circuit design, simulation, and optimization within a CAD simulator. Design and realization of a coplanar lumped element band pass filter on GaAs using the developed CPW EC-ANN models are demonstrated.

Nonlinear Modeling and Neuro-Fuzzy Control of PEMFC

The proton exchange membrane generation technology is highly efficient, and clea n and is considered as the most hopeful “green” power technology. The operatin g principles of proton exchange membrane fuel cell (PEMFC) system involve thermody namics, electrochemistry, hydrodynamics and mass transfer theory, which comprise a complex nonlinear system, for which it is difficult to establish a mathematic al model and control online. This paper analyzed the characters of the PEMFC; an d used the approach and self-study ability of artificial neural networks to bui ld the model of nonlinear system, and adopted the adaptive neural-networks fuzz y infer system to build the temperature model of PEMFC which is used as the refe rence model of the control system, and adjusted the model parameters to control online. The model and control were implemented in SIMULINK environment. The resu lts of simulation show the test data and model have a good agreement. The model is useful for the optimal and real time control of PEMFC system.

Identification and novel adaptive fuzzy control of nonlinear system for PEMFC stack

The operating temperature of a proton exchange membrane fuel cell stack is a very important control parameter. It should be controlled within a specific range, however, most of existing PEMFC mathematical models are too complicated to be effectively applied to on-line control. In this paper, input-output data and operating experiences will be used to establish PEMFC stack model and operating temperature control system. An adaptive learning algorithm and a nearest-neighbor clustering algorithm are applied to regulate the parameters and fuzzy rules so that the model and the control system are able to obtain higher accuracy. In the end, the simulation and the experimental results are presented and compared with traditional PID and fuzzy control algorithms.

Statistical Models for Long-range Forecasting of Southwest Monsoon Rainfall over India Using Step Wise Regression and Neural Network

The long-range forecasts (LRF) based on statistical methods for southwest monsoon rainfall over India (ISMR) has been issued by the India Meteorological Department (IMD) for more than 100 years. Many statistical and dynamical models including the operational models of IMD failed to predict the operational models of IMD failed to predict the deficient monsoon years 2002 and 2004 on the earlier occasions and so had happened for monsoon 2009. In this paper a brief of the recent methods being followed for LRF that is 8-parameter and 10-parameter power regression models used from 2003 to 2006 and new statistical ensemble forecasting system are explained. Then the new three stage procedure is explained. In this the most pertinent predictors are selected from the set of all the potential predictors for April, June and July models. The model equations are developed by using the linear regression and neural network techniques based upon training set of the 43 years of data from 1958 to 2000. The skill of the models is evaluated based upon the validation set of 11 years of data from 2001 to 2011, which has shown the high skill on the validation data set. It can be inferred that these models have the potential to provide a prediction of ISMR, which would significantly improve the operational forecast.

Prediction of Cracking Gas Compressor Performance and Its Application in Process Optimization

Cracking gas compressor is usually a centrifugal compressor. The information on the performance of a centrifugal compressor under all conditions is not available, which restricts the operation optimization for compressor. To solve this problem, two back propagation (BP) neural networks were introduced to model the performance of a compressor by using the data provided by manufacturer. The input data of the model under other conditions should be corrected according to the similarity theory. The method was used to optimize the system of a cracking gas compressor by embedding the compressor performance model into the ASPEN PLUS model of compressor. The result shows that it is an effective method to optimize the compressor system.

OPTIMAL PWM BASED ON REAL—TIME SOLUTION WITH NEURAL NETWORK

A novel concept of neural network based control in pulse-width modulation(PWM)voltage source inverters is presented.On the one hand,the optimal switching an-gles are obtained in real time by the neural network based controller；on the other hand,the output voltage is ad-justed to fit the expected value by neural network when input voltage or loads change.The structure of neural network is simple and easy to be realized by DSP hard-ware system.No large memory used for the existing opti-mal PWM schemes is required in the system.Theoreticalanlysis of the proposed so-called sparse neural network is provided,and the stability of the system is proved.Un-der the control of neural network the error of output volt-age descends sharply,and the system outputs ac voltage with high precision.

Simulation of CIECAM02 color appearance model based on Chinese color system

Color appearance model （CAM） can be used to determine the required colors for repro- duction across changes in cross-media. CIECAM02 color appearance model prediction is implemen- ted by artificial neural networks in this paper, which includes forward and reversed prediction. 1333 color samples as training samples and other 1332 color samples as test samples are selected in the Chinese color system. In order to test the prediction accuracy of neural networks after simulation of CIECAM02 color appearance model, the color-difference formula can be used for the evaluation of forward and reversed models. Results have shown that BP neural-network has acceptable accuracy in simulation of CIECAM02 color appearance model for colors of Chinese color system.

Quadratic stabilization of a nonlinear aeroelastic system using a novel Neural-Network-based controller

This contribution proposes a novel neural-network-based control approach to stabilize a nonlinear aeroelastic wing section. With the prerequisite that all the states of the system are available, the proposed controller requires no comprehensive information about structural nonlinearity of the wing section. Furthermore, the proposed control approach requires no human intervention of designing goal dynamics and formulating control input function, which is difficult to be realized by the typical neural-network-based control following an inverse control scheme. Simulation results show that the proposed controller can stabilize the aeroelastic system with different nonlinearities.