Related Experiment Videos
Identification of time-varying Hammerstein systems from ensemble data.
1Department of Biomedical Engineering, McGill University, Montréal, Québec, Canada.
Annals of Biomedical Engineering
|August 15, 2001
Summary
This study introduces a novel method for identifying fast-changing Hammerstein systems using input-output data. The technique combines correlation analysis and iterative optimization for accurate parameter estimation.
Area of Science:
- Systems Engineering
- Signal Processing
- Control Theory
Background:
- Hammerstein systems are widely used to model nonlinear dynamic processes.
- Identifying rapidly time-varying systems presents significant challenges.
- Existing methods often require specific input signal properties, such as white noise.
Purpose of the Study:
- To develop a new technique for identifying rapidly time-varying Hammerstein systems.
- To enable system identification without requiring white noise inputs.
- To provide accurate parameter estimates for complex dynamic systems.
Main Methods:
- A two-step approach combining correlation analysis and iterative optimization.
- Initial estimation of linear subsystem parameters using correlation.
- Refinement of system parameter estimates via an iterative optimization algorithm.
- The method accommodates non-white input signals.
Main Results:
- The proposed technique successfully identified rapidly time-varying Hammerstein systems.
- Excellent results were achieved on simulated data under realistic conditions.
- The method demonstrated robustness and accuracy in parameter estimation.
Conclusions:
- The developed technique offers an effective solution for identifying complex Hammerstein systems.
- This method advances the field of system identification for dynamic nonlinear processes.
- The approach is suitable for practical applications with non-ideal input signals.