Nonlinear system identification using optimized dynamic neural network

Popular Wiener models are a combination of dominant linear behavior and static nonlinear characteristics. However, they are less accurate in dealing with problems with strong nonlinear time-delayed effects. Hence, a group of novel nonlinear models based on Wiener models are proposed in this study. The static nonlinear part of general Wiener models is replaced by a time-delayed nonlinear block. Thus, two novel Wiener models with a quasi-dynamic nonlinear block are constructed, called the Wiener-QDN models. The autoregressive with exogenous input (ARX) model is utilized to identify the dynamic linear block, and the radial basis function neural network is used to build the nonlinear part of Wiener and the Wiener-QDN models. Results reveal the effectiveness of these models in nonlinear system identification. The first test cases are systems governed by simple equations. The Wiener-QDN models can identify these systems with or without measurement noise. These models are further applied in unsteady aerodynamic modeling cases. Test cases of a NACA 0012 airfoil pitching in transonic flow indicate that both linearity and nonlinearity are well captured by the proposed Wiener-QDN models.

show abstract

“…This testing input signal is similar to the input signal in [51], which is applied to validate the robustness of the three models.…”

Section: Examplesmentioning

confidence: 99%

Novel Wiener models with a time-delayed nonlinear block and their identification

2016

View full text Add to dashboard Cite

show abstract

“…Theorem 2: If we consider Sections II-A and B together, which means taking the identification and control as a whole process, we can apply the strategy to system (1).…”

Section: B Control Algorithmmentioning

confidence: 99%

“…Therefore, system identification becomes a relevant and necessary issue before system control can be considered. Recent research results show that dynamic neural networks (NNs) with different timescales are very effective for modeling the complex nonlinear systems with different timescales when we have incomplete model information, or even when we consider the plant as a black box [1]- [3].…”

Section: Introductionmentioning

confidence: 99%

Nonlinear Systems Identification and Control Via Dynamic Multitime Scales Neural Networks

Xie

Han

et al. 2013

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

This paper deals with the adaptive nonlinear identification and trajectory tracking via dynamic multilayer neural network (NN) with different timescales. Two NN identifiers are proposed for nonlinear systems identification via dynamic NNs with different timescales including both fast and slow phenomenon. The first NN identifier uses the output signals from the actual system for the system identification. In the second NN identifier, all the output signals from nonlinear system are replaced with the state variables of the NNs. The online identification algorithms for both NN identifier parameters are proposed using Lyapunov function and singularly perturbed techniques. With the identified NN models, two indirect adaptive NN controllers for the nonlinear systems containing slow and fast dynamic processes are developed. For both developed adaptive NN controllers, the trajectory errors are analyzed and the stability of the systems is proved. Simulation results show that the controller based on the second identifier has better performance than that of the first identifier.

show abstract

“…The control action uconsists of two parts: L f = + u u u (15) where L u is a compensation for the known nonlinearity and u f are dedicated to deal with the model error, which can be left open if it is zero or ignorable. Let L u be ( ) (16) and rewrite (14) as…”

Section: Tracking Error Analysismentioning

confidence: 99%

“…Therefore, system identification becomes a relevant issue and even necessary before system control can be considered. Recent research results show that dynamic neural networks with different time-scales are very effective for modeling the complex nonlinear systems with different time-scales when we have incomplete model information, or even when we consider the plant as a black-box [15] [17].…”

Section: Introductionmentioning

confidence: 99%

Indirect adaptive control of nonlinear system via dynamic multilayer neural networks with multi-time scales

Liu

Xie

2013

2013 Fourth International Conference on Intelligent Control and Information Processing (ICICIP)

View full text Add to dashboard Cite

This paper deals with the adaptive nonlinear identification and trajectory tracking via dynamic multilayer neural network with different time-scales. By means of a Lyapunov-like analysis we determine stability conditions for the identification. Based on the identification results, we design a sliding mode controller for the nonlinear system to track the trajectory of a reference model. The main contributions of the paper are: First, we extend our prior results of single-layer dynamic neural networks with multi-time scales to the multilayer case. Second, the e-modification in standard use in adaptive control is introduced in the on-line update laws to guarantee bounded weights, bounded identification and tracking errors. Simulation results are presented confirming the validity of the above approach.

show abstract

Nonlinear system identification using optimized dynamic neural network

Cited by 37 publications

References 14 publications

Novel Wiener models with a time-delayed nonlinear block and their identification

Novel Wiener models with a time-delayed nonlinear block and their identification

Nonlinear Systems Identification and Control Via Dynamic Multitime Scales Neural Networks

Indirect adaptive control of nonlinear system via dynamic multilayer neural networks with multi-time scales

Contact Info

Product

Resources

About