We propose a new analytic method for comparing constant gain adaptive signal processing algorithms. Specifically, estimates of the convergence speed of the algorithms allow for the definition of a local measure of performance (the Efficacy) that can be theoretically evaluated. By definition, the Efficacy is consistent with the fair comparison techniques introduced lately in the literature (based on simulations) and can consequently be used as a theoretical alternative to these methods. Using the Efficacy as a performance measure, we prove that the NewtonLMS algorithm is optimum and is thus the fastest algorithm in a very rich algorithmic class. Furthermore, we prove that the regular LMS is better than any of its variants, which apply a nonlinear transformation on the elements of its regression vector (such as signed régresser, quantized régresser, etc.) for an important class of input signals. Simulations support all our theoretical developments.
|Original language||English (US)|
|Number of pages||1|
|Journal||IEEE Transactions on Signal Processing|
|State||Published - 1998|
All Science Journal Classification (ASJC) codes
- Signal Processing
- Electrical and Electronic Engineering