Characterization of nonlinear input-output systems using time series analysis
Data obtained from time series analysis has been used for a number of years for the characterization and response prediction of linear systems. This paper describes a time series technique for the analysis of nonlinear systems through the use of embeddings using delay coordinates or appropriate transformations of delay coordinates (local singular value decomposition or local canonical variate analysis). Local linear approaches are used to characterize the state evolution. Application of the technique is illustrated for a single degree of freedom oscillator with nonlinear stiffness, a mechanical chaos beam, and a climatic data time series. In each application analysis from measured data is emphasized. State rank, lyapunov exponents, and expected iterated prediction errors are quantified. The technique illustrated should be useful in the analysis of many forms of experimental data, especially where the state rank is not excessively large. 8 refs., 6 figs.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE; DOHHS; USDOE, Washington, DC (United States); Department of Health and Human Services, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 5360673
- Report Number(s):
- LA-UR-91-2816; CONF-9110200-1; ON: DE91018003
- Resource Relation:
- Conference: 1. experimental chaos conference, Arlington, VA (United States), 1-3 Oct 1991
- Country of Publication:
- United States
- Language:
- English
Similar Records
Using surrogate data to detect nonlinearity in time series
Nonlinear dynamical modeling and prediction of the terrestrial magnetospheric activity