The advantages of directly identifying continuous-time transfer function models in practical applications

Abstract

The direct identification and estimation of continuous-time models from sampled data is now mature. This paper does not present any new methodology, nor does it compare the performance of existing methods. Its main aim is to discuss the advantages of direct, continuous-time model identification with the help of illustrative examples that are all based on real data from practical applications. Although the specific method of statistical parameter estimation is relatively unimportant in this regard, the latest and most sophisticated time domain identification algorithm that is freely available to the reader is used in these examples in order to ensure that the results reflect the best performance that can be achieved at this time by time-domain identification.

Keywords:

• continuous-time model
• discrete-time data
• instrumental variable
• real-life data
• system identification

Acknowledgements

The authors are grateful to anonymous referees whose comments helped to clarify and improve the paper.

Notes

1. Note however that DT models with fractional time-delay can easily be estimated in the frequency domain (see, for example Pintelon and Schoukens (2001)).

2. This can be freely downloaded from www.cran.uhp-nancy.fr/contsid/ and includes many demonstration examples (Garnier, Gilson, Bastogne, & Mensler, 2008).

3. This can be freely downloaded from www.es.lancs.ac.uk/cres/captain/ and includes many demonstration examples. It is also supported by both a handbook (Pedregal, Taylor, & Young, 2005) and a recently published textbook (Young, 2011).

4. This is not essential to the functioning of the algorithm: it allows for rarely occurring situations, normally with very poor data, when the initially estimated model is found to be unstable.

5. The authors are most grateful to Prof. de Callafon from UCSD for kindly providing us with these data.

6. The authors gratefully acknowledge Prof. de Callafon and the Undergraduate and Graduate Control Laboratory (UGCL) at UCSD for providing the data.