Indirect inverse adaptive control based on neural networks using dynamic back propagation for nonlinear dynamic systems

The paper proposes a fast online learning method for neural network structures by using genetic algorithm (GA) and dynamic back propagation algorithm (BP) jointly. GA is used in the coarse tuning process which adjusts interconnection weights of the neural network. The dynamic back propagation algorithm is subsequently applied to achieve fine adjusting of the network weights. The fitness function, based on the squared error between the teaching signal and the network output value, is redefined at every time step and the proposed GA based algorithm solves a nonstationary function optimization task. At every time step the solution with the best fitness function is used for current representation of the neural network weights and biases. It is shown through the simulations and real time temperature control of drying oven that this learning algorithm has faster convergence ability and better performance on reducing mapping error in the online learning neural network structures. This leads to an improvement of the transient response of neuro adaptive systems. The proposed method has the potential to be applied to many practical areas such as system modeling and control, signal processing and pattern recognition.

Among all of the Machine Learning techniques used for classification tasks, Artificial Neural Networks (ANNs) have obtained much success in their applications. However, their development usually requires a manual effort from the human expert in which several parameter configurations (architectures, training parameters, etc) are tried. This paper proposes a new evolutionary method that evolves ANNs without any participation from the human expert. This system can be used to evolve feed-forward and recurrent ANNs. A real-world problem has been used to test the behaviour of this system: detection of epileptic seizures in EEG signals. A comparison of the results obtained using recurrent and feedforward ANNs to solve this problem is presented in this paper. This comparison shows the good accuracies obtained by this method (almost 100%). Moreover, these results show an important feature: the system tries to evolve simple ANNs, with a low number of neurons and connections (in many cases, the networks have only 1 hidden neuron).

A new hybrid recurrent neural network (HRNN) for machining process modeling is presented based on the diagonal recurrent neural network (DRNN). In order to overcome the weakness of back propagation (BP) algorithm, a generalized entropy square error (GESE) criterion is defined and a dynamic recurrent back propagation algorithm is developed to guarantee the global convergence. The HRNN based on the GESE is then used for nonlinear system identification and neural network modeling of the machining process. The numeral experiments results show that the HRNN has better approximate effectiveness, tracking and dynamic performance than traditional BP neural network.

Artificial neural networks (ANNs) are used to determine the state-of-health (SOH) of the Automated Radioxenon Analyzer/Sampler (ARSA). ARSA is a gas collection and analysis system used for non-proliferation monitoring in detecting radioxenon released during nuclear tests. SOH diagnostics are important for automated, unmanned sensing systems so that remote detection and identification of problems can be made without onsite staff. Both recurrent and feed-forward ANNs are presented. The recurrent ANN is trained to predict sensor values based on current valve states, which control air flow, so that with only valve states the normal SOH sensor values can be predicted. Deviation between modeled value and actual is an indication of a potential problem. The feed-forward ANN acts as a nonlinear version of principal components analysis (PCA) and is trained to replicate the normal SOH sensor values. Because of ARSA's complexity, this nonlinear PCA is better able to capture the relationships among the sensors than standard linear PCA and is applicable to both sensor validation and recognizing off-normal operating conditions. Both models provide valuable information to detect impending malfunctions before they occur to avoid unscheduled shutdown. Finally, the ability of ANN methods to predict the system state is presented.

Brain research up to date has revealed that structure and function are highly related. Thus, for example, studies have repeatedly shown that the brains of patients suffering from schizophrenia or other diseases have a different connectome compared to healthy people. Apart from stochastic processes, however, an inherent logic describing how neurons connect to each other has not yet been identified. We revisited this structural dilemma by comparing and analyzing artificial and biological-based neural networks. Namely, we used feed-forward and recurrent artificial neural networks as well as networks based on the structure of the micro-connectome of C. elegans and of the human macro-connectome. We trained these diverse networks, which markedly differ in their architecture, initialization and pruning technique, and we found remarkable parallels between biological-based and artificial neural networks, as we were additionally able to show that the dilemma is also present in artificial neural networks. Our findings show that structure contains all the information, but that this structure is not exclusive. Indeed, the same structure was able to solve completely different problems with only minimal adjustments. We particularly put interest on the influence of weights and the neuron offset value, as they show a different adaption behaviour. Our findings open up new questions in the fields of artificial and biological information processing research.

Despite their ubiquity, Artificial Neural Networks (ANNs) are still difficult to correctly implement and verify. While feedforward ANNs behave based on a clear cause and effect that maps an input to an output, Recurrent ANNs (RANNs) propagate internal (i.e., hidden layer) or external (i.e., output layer) values back to the inputs. Problematically, the benefits enabled by the recurrent links also introduce verification challenges. The cause and effect relationship of input to output units normally displayed in feedforward artificial neural networks may be altered due to internal values propagated forward in time via recurrent links. This verification difficulty is exacerbated in embedded systems (e.g., robotic or cyber-physical systems software) where in-house implementations of RANNs may be preferred, or even necessitated, due to hardware constraints. In this paper, we identify minimalistic test cases to verify the behavior of context units connected by internal recurrent links within RANNs.

This paper presents how to apply a newly developed general theory for the global transversality and tangency of flows in n-dimensional nonlinear dynamical systems to a 2-D nonlinear dynamical system (i.e. a periodically forced, damped Duffing oscillator). The global tangency and transversality of the periodic and chaotic motions to the separatrix for such a nonlinear system are discussed to help us understand the complexity of chaos in nonlinear dynamical systems. This paper presents the concept that the global transversality and tangency to the separatrix are independent of the Melnikov function (or the energy increment). Chaos in nonlinear dynamical systems makes the exact energy increment quantity to be chaotic no matter if the nonlinear dynamical systems have separatrices or not. The simple zero of the Melnikov function cannot be used to simply determine the existence of chaos in nonlinear dynamical systems. Through this paper, the expectation is that, from now on, one can use the alternative aspect to look into the complexity of chaos in nonlinear dynamical systems. Therefore, in this paper, the analytical conditions for global transversality and tangency of 2-D nonlinear dynamical systems are presented. The first integral quantity increment (i.e. the energy increment) for a certain time interval is achieved for periodic flows and chaos in the 2-D nonlinear dynamical systems. Under the perturbation assumptions and convergent conditions, the Melnikov function is recovered from the first integral quantity increment. A periodically forced, damped Duffing oscillator with a separatrix is investigated as a sampled problem. The corresponding analytical conditions for the global transversality and tangency to the separatrix are obtained and verified by numerical simulations. The switching planes and the corresponding local and global mappings are defined on the separatrix. The mapping structures are developed for local and global periodic flows passing through the separatrix. The mapping structures of global chaos in the damped Duffing oscillator are also discussed. Bifurcation scenarios of the damped Duffing oscillator are presented through the traditional Poincaré mapping section and the switching planes. The first integral quantity increment (i.e. L-function) is presented to observe the periodicity of flows. In addition, the global tangency of periodic flows in such an oscillator is measured by the G-function and G(1)-function, and is verified by numerical simulations. The first integral quantity increment of periodic flows is zero for their complete periodic cycles. Numerical simulations of chaos in such a Duffing oscillator are carried out through the Poincaré mapping sections. The conservative energy distribution, G-function and L-function along the displacement of Poincaré mapping points are presented to observe the complexity of chaos. The first integral quantity increment (i.e. L-function) of chaotic flows at the Poincaré mapping points is nonzero and chaotic. The switching planes of chaos are presented on the separatrix for a better understanding of the global transversality to the separatrix. The switching point distribution on the separatrix is presented and the switching G-function on the separatrix is given to show the global transversality of chaos on the separatrix. The analytical conditions are obtained from the new theory rather than the Melnikov method. The new conditions for the global transversality and tangency are more accurate and independent of the small parameters.

Ecological Applications of Fuzzy Logic.- Ecological Applications of Qualitative Reasoning.- Ecological Applications of Non-supervised Artificial Neural Networks.- Ecological Applications of Genetic Algorithms.- Ecological Applications of Evolutionary Computation.- Ecological Applications of Adaptive Agents.- Bio-Inspired Design of Computer Hardware by Self-Replicating Cellular Automata.- Prediction and Elucidation of Stream Ecosystems.- Development and Application of Predictive River Ecosystem Models Based on Classification Trees and Artificial Neural Networks.- Modelling Ecological Interrelations in Running Water Ecosystems with Artificial Neural Networks.- Non-linear Approach to Grouping, Dynamics and Organizational Informatics of Benthic Macroinvertebrate Communities in Streams by Artificial Neural Networks.- Elucidation of Hypothetical Relationships between Habitat Conditions and Macroinvertebrate Assemblages in Freshwater Streams by Artificial Neural Networks.- Prediction and Elucidation of River Ecosystems.- Prediction and Elucidation of Population Dynamics of the Blue-green Algae Microcystis aeruginosa and the Diatom Stephanodiscus hantzschii in the Nakdong River-Reservoir System (South Korea) by a Recurrent Artificial Neural Network.- An Evaluation of Methods for the Selection of Inputs for an Artificial Neural Network Based River Model.- Utility of Sensitivity Analysis by Artificial Neural Network Models to Study Patterns of Endemic Fish Species.- Prediction and Elucidation of Lake and Marine Ecosystems.- A Comparison between Neural Network Based and Multiple Regression Models for Chlorophyll-a Estimation.- Artificial Neural Network Approach to Unravel and Forecast Algal Population Dynamics of Two Lakes Different in Morphometry and Eutrophication.- Hybrid Evolutionary Algorithm for Rule Set Discovery in Time-Series Data to Forecast and Explain Algal Population Dynamics in Two Lakes Different in Morphometry and Eutrophication.- Multivariate Time Series Prediction of Marine Zooplankton by Artificial Neural Networks.- Classification of Fish Stock-Recruitment Relationships in Different Environmental Regimes by Fuzzy Logic with Bootstrap Re-sampling Approach.- Computational Assemblage of Ordinary Differential Equations for Chlorophyll-a Using a Lake Process Equation Library and Measured Data of Lake Kasumigaura.- Classification of Ecological Images at Micro and Macro Scale.- Identification of Marine Microalgae by Neural Network Analysis of Simple Descriptors of Flow Cytometric Pulse Shapes.- Age Estimation of Fish Using a Probabilistic Neural Network.- Pattern Recognition and Classification of Remotely Sensed Images by Artificial Neural Networks.

Unlimited access to information and data sharing wherever and at any time for anyone and anything is a fundamental component of fifth-generation (5G) wireless communication and beyond. Therefore, it has become inevitable to exploit the super-high frequency (SHF) and millimeter-wave (mmWave) frequency bands for future wireless networks due to their attractive ability to provide extremely high data rates because of the availability of vast amounts of bandwidth. However, due to the characteristics and sensitivity of wireless signals to the propagation effects in these frequency bands, more accurate path loss prediction models are vital for the planning, evaluating, and optimizing future wireless communication networks. This paper presents and evaluates the performance of several well-known machine learning methods, including multiple linear regression (MLR), polynomial regression (PR), support vector regression (SVR), as well as the methods using decision trees (DT), random forests (RF), K-nearest neighbors (KNN), artificial neural networks (ANN), and artificial recurrent neural networks (RNN). RNNs are mainly based on long short-term memory (LSTM). The models are compared based on measurement data to provide the best fitting machine-learning-based path loss prediction models. The main results obtained from this study show that the best root-mean-square error (RMSE) performance is given by the ANN and RNN-LSTM methods, while the worst is for the MLR method. All the RMSE values for the given learning techniques are in the range of to dB. Furthermore, this work shows that the models (except for the MLR model) perform excellently in fitting actual measurement data for wireless communications in enclosed indoor environments since they provide R-squared and correlation values higher than and , respectively. The paper shows that these learning methods could be used as accurate and stable models for predicting path loss in the mmWave frequency regime.

SummaryThis research article presents an artificial neural network (ANN)‐based indirect adaptive control method for nonlinear dynamical systems. In this article, a modified Elman recurrent neural network (MERNN) is proposed as an identifier and controller for controlling nonlinear systems. The architecture of the proposed controller is a modified form of the existing Elman recurrent neural network. The parameter training of ANN‐based controllers is obtained by using the most popular optimization algorithm which is known as the back‐propagation algorithm. A comparative study includes Elman, Diagonal, Jordan, feed‐forward neural network (FFNN), and radial basis function network (RBFN)‐based controllers to compare with the proposed MERNN controller. To determine the controller's robustness, parameter variations, and disturbance signals have been considered. The performance analysis of the proposed controller is illustrated by two simulation examples. The simulation results reveal that MERNN can not only identify the unknown dynamics of the plant but also adaptively control it compared to the others.

This research frame work investigates the application of a clustered based Neuro-fuzzy system to nonlinear dynamic system modeling from a set of input-output training patterns. It is concentrated on the modeling via Takagi-Sugeno (T-S) modeling technique and the employment of fuzzy clustering to generate suitable initial membership functions. Hence, such created initial memberships are then employed to construct suitable T-S sub-models. Furthermore, the T-S fuzzy models have been validated and checked through the use of some standard model validation techniques (like the correlation functions). Compared to other well-known approximation techniques such as artificial neural networks, fuzzy systems provide a more transparent representation of the system under study, which is mainly due to the possible linguistic interpretation in the form of rules. Such intelligent modeling scheme is very useful once making complicated systems linguistically transparent in terms of fuzzy if-then rules. The developed T-S Fuzzy modeling system has been then applied to model a nonlinear antenna dynamic system with two coupled inputs and outputs. Validation results have resulted in a very close antenna sub-models of the original nonlinear antenna system. The suggested technique is very useful for development transparent linear control systems even for highly nonlinear dynamic systems.

The goal of this paper is to introduce a new neural network architecture called Sigmoid Diagonal Recurrent Neural Network (SDRNN) to be used in the adaptive control of nonlinear dynamical systems. This is done by adding a sigmoid weight victor in the hidden layer neurons to adapt of the shape of the sigmoid function making their outputs not restricted to the sigmoid function output. Also, we introduce a dynamic back propagation learning algorithm to train the new proposed network parameters. The simulation results showed that the (SDRNN) is more efficient and accurate than the DRNN in both the identification and adaptive control of nonlinear dynamical systems.

Temporally local recurrent radial basis function network for modeling and adaptive control of nonlinear systems

This article presents static and recurrent artificial neural networks (ANNs) to predict the drying kinetics of carrot cubes during fluidized bed drying. Experiments were performed on square-cubed carrot with dimensions of 4, 7 and 10mm, air temperatures of 50, 60 and 70°C and bed depths of 3, 6 and 9cm. Initially, static ANN was used to correlate the outputs (moisture ratio and drying rate) to the four exogenous inputs (drying time, drying air temperature, carrot cubes size, and bed depth). In the recurrent ANNs, in addition to the four exogenous inputs, two state input and output (moisture ratio or drying rate) were applied. A number of hidden neurons and training epoch were investigated in this study. The dying kinetics was predicted with R(2) values of greater than 0.94 and 0.96 using static and recurrent ANNs, receptively.

Novel recurrent neural network for modelling biological networks: Oscillatory p53 interaction dynamics