Weak convergence and asymptotic properties of adaptive filters with constant gains
โ Scribed by Kushner, H.; Shwartz, A.
- Book ID
- 114635388
- Publisher
- IEEE
- Year
- 1984
- Tongue
- English
- Weight
- 706 KB
- Volume
- 30
- Category
- Article
- ISSN
- 0018-9448
No coin nor oath required. For personal study only.
๐ SIMILAR VOLUMES
We consider the almost sure convergence of the adaptive IIR filter based on the output error method. It is rigorously shown that the algorithm is globally stable, parameter estimation is consistent, and perfect noise cancellation is achieved in the presence of nonstationary colored noise. Due to the
Ahstraet--A modified version of Fortescue's adaptive regulator is described, based on a minimum-variance estimator with variable forgetting factor and a d-step ahead control law. As in the Fortescue algorithm, the forgetting factor is chosen at each step to keep a measure of the information content