Abstract
This paper proposes a new approach to adaptive identification of linear noisy input-output systems. This adaptive parameter estimation algorithm is developed based on the bias-correction principle, thus having the expected consistent convergence. Moreover, this algorithm does not involve prefiltering of noisy measurements or parameter extraction during the adaptation. Computer simulations are used to demonstrate its good performance in comparison with other on-line identification methods.
| Original language | English |
|---|---|
| Pages (from-to) | 474-488 |
| Number of pages | 15 |
| Journal | Dynamics of Continuous, Discrete and Impulsive Systems, Series B: Applications and Algorithms |
| Volume | 6 |
| Issue number | 4 |
| Publication status | Published - Dec 1999 |