Identification of Chaotic Systems by Neural Networks
Barbara Cannas2012, Chaos and Complex Systems
Sign up for access to the world's latest research
checkGet notified about relevant papers
checkSave papers to use in your research
checkJoin the discussion with peers
checkTrack your impact
Abstract
In this paper a traditional Multi Layer Perceptron with a tapped delay line as input is trained to identify the parameters of the Chua's circuit when fed with a sequence of values of a scalar state variable. The analysis of the a priori identifiability of the system, performed resorting to differential algebra, allows one to choose a suitable observable and the minimum number of taps. The results confirm the appropriateness of the proposed approach.
Key takeaways
- In this paper, a neural network is applied to parameter identification of chaotic systems starting from suitable observable variables.
- Parameter identification concerns the problem of which parameters of an ordinary differential equation (ODE) model can be evaluated from a given output.
- The Multi Layer Perceptron (MLP) is the most widely used type of neural network.
- When α=10, β=16, c=-0.143, Chua's circuit exhibit a chaotic behaviour.
- Then, chosen a proper observable, a neural network has been trained to learn the relationship between a sequence of the observable values and one unknown parameter.
Related papers
Projectional dynamic neural network identifier for chaotic systems: Application to Chua’s circuit
2011
Chaotic systems arise in many different fields, being their modeling, synchronization and control, important topics for interested researchers in this kind of systems. The present paper deals with the problem of non-parametric identification (adaptable modeling) of a class of uncertain nonlinear systems with chaotic behavior. A Projectional Dynamic Neural Network (PDNN) is proposed to carry out the identification task. The practical stability of the identification error is proven by the second Lyapunov's method and the Linear Matrix Inequalities approaches. The obtained algorithm is tested by numerical simulations, taking into account the mathematical model of the so-called chaotic Chuas circuit, and compared with a nonlinear observer (Thaus observer). Then, a reported version of the Chua's circuit is constructed to verify the proposed identification scheme under an experimental framework. PDNN is tested with this real data. In both cases, simulation and real measurements, the developed algorithm shows an excellent convergence to the state variables of the chaotic system, fact that is supported by the formal analysis.
Neural networks for signal processing measurement analysis and industrial applications: the case of chaotic signal processing
This chapter discusses the use of neural networks for signal processing. In particular, it focuses on one of the most interesting and innovative areas: the chaotic time series processing. This includes time series analysis, identification of chaotic behavior, forecasting, and dynamic reconstruction. An overview of chaotic signal processing both by conventional and neural network methods is given.
Identification of chaotic systems by neural network with hybrid learning algorithm
Chaos, Solitons & Fractals, 2008
Based on the genetic algorithm (GA) and steepest descent method (SDM), this paper proposes a hybrid algorithm for the learning of neural networks to identify chaotic systems. The systems in question are the logistic map and the Duffing equation. Different identification schemes are used to identify both the logistic map and the Duffing equation, respectively. Simulation results show that our hybrid algorithm is more efficient than that of other methods.
Identification and control of unknown chaotic systems via dynamic neural networks
IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications, 1999
Identification and control problems for unknown chaotic dynamical systems are considered. Our aim is to regulate the unknown chaos to a fixed point or a stable periodic orbit. This is realized by following two contributions. First, a dynamic neural network is used as identifier. The weights of the neural networks are adjusted by the sliding mode technique. Second, we derive a local optimal controller via the neuroidentifier to remove the chaos in a system. The identification error and trajectory error are guaranteed to be bounded. The controller proposed in this paper is effective for many chaotic systems, including the Lorenz system, Duffing equation, and Chua's circuit.
Pattern recognition using chaotic neural networks
Discrete Dynamics in Nature and Society, 1998
Pattern recognition by chaotic neural networks is studied using a hyperchaotic neural network as model. Virtual basins of attraction are introduced around unstable periodic orbits which are then used as patterns. Search for periodic orbits in dynamical systems is treated as a process of pattern recognition. The role of synapses on patterns in chaotic networks is discussed. It is shown that distorted states having only limited information of the patterns are successfully recognized.
Parametric Identification of Nonlinear Systems Using Chaotic Excitation
Journal of Computational and Nonlinear Dynamics, 2007
Parametric identification of a single degree-of-freedom (SDOF) nonlinear Duffing oscillator is carried out using a harmonic balance (HB) method. The parameters of the system are obtained using a harmonic input, for the case of periodic response. Problems of matrix inversion, due to poor conditioning are sometimes encountered in the computation. This may occur due to large differences in the relative values of inertia, damping and spring forces or dependence of these parameters on one another. The inversion problem may also occur due to a poor choice of the excitation signal frequency and amplitude. However there is limited choice of adjustable input parameters in this case. In this work, an extended HB method, which uses a combination of two harmonic inputs, is suggested to overcome the above problem.
Neural learning of chaotic dynamics: the error propagation algorithm
1998 IEEE International Joint Conference on Neural Networks Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98CH36227)
An algorithm is introduced that trains a neural network to identify chaotic dynamics from a single measured timeseries. The algorithm has four special features: 1. The state of the system is extracted from the time-series using delays, followed by weighted Principal Component Analysis (PCA) data reduction. 2. The prediction model consists of both a linear model and a Multi-Layer-Perceptron (MLP). 3. The effective prediction horizon during training is user-adjustable, due to 'error propagation': prediction errors are partially propagated to the next time step. 4. A criterion is monitored during training to select the model that has a chaotic attractor most similar to the real system's attractor. The algorithm is applied to laser data from the Santa Fe time-series competition (set A). The resulting model is not only useful for short-term predictions but it also generates time-series with similar chaotic characteristics as the measured data.
New artificial neural network design for Chua chaotic system prediction using FPGA hardware co-simulation
International Journal of Electrical and Computer Engineering (IJECE), 2022
This study aims to design a new architecture of the artificial neural networks (ANNs) using the Xilinx system generator (XSG) and its hardware co-simulation equivalent model using field programmable gate array (FPGA) to predict the behavior of Chua's chaotic system and use it in hiding information. The work proposed consists of two main sections. In the first section, MATLAB R2016a was used to build a 3×4×3 feed forward neural network (FFNN). The training results demonstrate that FFNN training in the Bayesian regulation algorithm is sufficiently accurate to directly implement. The second section demonstrates the hardware implementation of the network with the XSG on the Xilinx artix7 xc7a100t-1csg324 chip. Finally, the message was first encrypted using a dynamic Chua system and then decrypted using ANN's chaotic dynamics. ANN models were developed to implement hardware in the FPGA system using the IEEE 754 Single precision floating-point format. The ANN design method illustrated can be extended to other chaotic systems in general.
Modelling concepts arising from an investigation into a chaotic system
Mathematical and Computer Modelling, 1991
The Chua circuit, a simple third-order nonlinear dynamical system that exhibits chaotic behaviour, has highlighted shortcomings in presently used analysis and design techniques. We discuss the way in which chaotic dynamical behaviour in systems has affected many current ideas about the randomness of noise, the use of simulation, reducing problems to subproblems, the value of experiments, the use of nonlinear controllers, aud the verification of system models.
Cryptography based on Artificial Neural Networks and Chaos Theory
International Journal of Computer Applications, 2016
Cryptography is a skill of sending the data in such a form that only those for whom it is intended can read it. There are number of methods to perform cryptography, one of such methods is Chaos theory which studies the behavior of a dynamical systems that are highly sensitive to initial conditions. Even slight changes in initial conditions result in extensively deviating outcomes for such dynamical systems, hence making long-standing estimate unmanageable. The limitations of applying Chaos Theory are choosing the input parameters and synchronization. The computation of these input parameters lies on the dynamics underlying the data and the highly complex analysis, not always accurate. Artificial neural networks (ANN) well known for learning and generalization are hence used to model the dynamics of Chua's circuit viz. x, y and z. The designed ANN was trained by varying its structures and using different learning algorithms. ANN was trained using 9 different sets which were formed with the initial conditions of Chua's circuit and each set consisted of about 1700 input-output data. A feedforward Multi-Layer Perceptron (MLP) network structure, trained with Levenberg-Marquardt backpropagation algorithm, produced best outcome. Further a case study in which a plain text was first encoded and then decoded by both the chaotic dynamics obtained from the proposed ANN and the numerical solution of Chua's circuit and are compared with each other.
Related papers
Neural networks and chaos: Construction, evaluation of chaotic networks, and prediction of chaos with multilayer feedforward networks
Chaos: An Interdisciplinary Journal of Nonlinear Science, 2012
Many research works deal with chaotic neural networks for various fields of application. Unfortunately, up to now these networks are usually claimed to be chaotic without any mathematical proof. The purpose of this paper is to establish, based on a rigorous theoretical framework, an equivalence between chaotic iterations according to Devaney and a particular class of neural networks. On the one hand we show how to build such a network, on the other hand we provide a method to check if a neural network is a chaotic one. Finally, the ability of classical feedforward multilayer perceptrons to learn sets of data obtained from a dynamical system is regarded. Various Boolean functions are iterated on finite states. Iterations of some of them are proven to be chaotic as it is defined by Devaney. In that context, important differences occur in the training process, establishing with various neural networks that chaotic behaviors are far more difficult to learn.
The Modeling of the Rucklidge Chaotic System with Artificial Neural Networks
Chaos Theory and Applications
Chaotic systems are nonlinear systems that show sensitive dependence on initial conditions, and an immeasurably small change in initial value causes an immeasurably large change in the future state of the system. Besides, there is no randomness in chaotic systems and they have an order within themselves. Researchers use chaotic systems in many areas such as mixer systems that can make more homogeneous mixtures, encryption systems that can be used with high security, and artificial neural networks by taking the advantage of the order in this disorder. Differential equations in which chaotic systems are expressed mathematically are solved by numerical solution methods such as Heun, Euler, ODE45, RK4, RK5-Butcher and Dormand-Prince in the literature. In this research, Feed Forward Neural Network (FFNN), Layer Recurrent Neural Network (LRNN) and Cascade Forward Backpropogation Neural Network (CFNN) structures were used to model the Rucklidge chaotic system by making use of the MATLAB R2...
Reconstruction of chaotic time series by neural models: a case study
Neurocomputing, 2003
This work analyses the problems related to the reconstruction of a dynamical system, which exhibits chaotic behaviour, from time series associated with a single observable of the system itself, by using feedforward neural network model. The starting network architecture is obtained setting the number of input neurons according to the Takens' theorem, and then is imporved by slightly increasing the number of inputs. The choice of the number of the hidden neurons is based on the results obtained testing di erent net structures. The e ectiveness of the method is demonstrated by applying it to the Brusselator system (Phys. Lett. 91 (1982) 263).
Parameter Identification of Chaotic Systems Using a Modified Cost Function Including Static and Dynamic Information of Attractors in the State Space
Circuits Systems and Signal Processing, 2018
A c c e p t e d M a n u s c r i p t Qiaoyong Jiang received the B.S. degree in mathematics from Wenzhou University, Wenzhou, China, in 2006, the M.S. degree in applied mathematics from Beifang University of Nationalities, Yinchuan, China, in 2011. He is currently pursuing the Ph.D. degree in pattern
Identifiability and identification of chaotic systems based on adaptive synchronization
IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications, 1997
This paper deals with the problem of synchronization of chaotic systems when the driven (slave, receiver) system has the same structure as the master (driving, emitter) system but its parameters are unknown. It is shown that the concept of synchronization provides an efficient way to find the unknown slave system parameters. Parameter mismatch between master and slave systems and high sensitivity of response to changes of these parameters were so far considered as crucial for security issues. This paper shows evidence that this claimed advantage becomes in fact a major drawback in chaos communication schemes since parameters can easily be found using adaptive synchronization and optimization tools. The general problem of identifiability of chaotic systems is defined and discussed in the context of possibilities for finding the unknown chaotic receiver parameters. Several typical systems used in experiments in chaos communication are tested for identifiability showing direct applications of the introduced concepts. In particular examples of the skew tent map, Hénon map, Markov maps and Chua's circuit are considered in detail illustrating the problems of global and local identifiability.
Chaotic time series part II: System identification and prediction
Arxiv preprint chao-dyn/ …, 1994
This paper is the second in a series of two, and describes the current state of the art in modelling and prediction of chaotic time series.
Prediction of Chaotic Time Series with Neural Networks and the Issue of Dynamic Modeling
International Journal of Bifurcation and Chaos, 1992
This paper shows that the dynamics of nonlinear systems that produce complex time series can be captured in a model system. The model system is an artificial neural network, trained with backpropagation, in a multi-step prediction framework. Results from the Mackey-Glass (D=30) will be presented to corroborate our claim. Our final intent is to study the applicability of the method to the electroencephalogram, but first several important questions must be answered to guarantee appropriate modeling.
Learning of Chua's circuit attractors by locally recurrent neural networks
2001
Many practical applications of neural networks require the identification of strongly non-linear (eg, chaotic) systems. In this paper, locally recurrent neural networks (LRNNs) are used to learn the attractors of Chua's circuit, a paradigm for studying chaos. LRNNs are characterized by a feed-forward structure whose synapses between adjacent layers have taps and feedback connections. In general, the learning procedures of LRNNs are computationally simpler than those of globally recurrent networks.
Evolutionary identification of chaotic system
Second IFAC Conference on Analysis and Control of Chaotic Systems (2009), 2009
Synthesis, identification and control of complex dynamical systems are usually extremely complicated. When classics methods are used, some simplifications are required which tends to lead to idealized solutions that are far from reality. In contrast, the class of methods based on evolutionary principles is successfully used to solve this kind of problems with a high level of precision. In this paper an alternative method of evolutionary algorithms, which has been successfully proven in many experiments like chaotic systems synthesis, neural network synthesis or electrical circuit synthesis. This paper discusses the possibility of using evolutionary algorithms for the identification of chaotic systems. The main aim of this work is to show that evolutionary algorithms are capable of the identification of chaotic systems without any partial knowledge of internal structure, i.e. based only on measured data. Two different evolutionary algorithms are presented and tested here in a total of 10 versions. Systems selected for numerical experiments here is the well-known logistic equation. For each algorithm and its version, repeated simulations were done, amounting to 50 simulations. According to obtained results it can be stated that evolutionary identification is an alternative and promising way as to how to identify chaotic systems.
Loading Preview
Sorry, preview is currently unavailable. You can download the paper by clicking the button above.