A new class of self‐normalising LMS algorithms
Abstract Many researchers and practitioners make heavy use of the least mean squares (LMS) algorithm as an efficient adaptive filter suitable for a multitude of problems. Despite being versatile and efficient, a drawback of this algorithm is that the adaptation rate, i.e. step‐size, has to be chosen...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2022-06-01
|
| Series: | Electronics Letters |
| Online Access: | https://doi.org/10.1049/ell2.12498 |
| _version_ | 1811159803621277696 |
|---|---|
| author | Oliver Ploder Oliver Lang Thomas Paireder Christian Motz Mario Huemer |
| author_facet | Oliver Ploder Oliver Lang Thomas Paireder Christian Motz Mario Huemer |
| author_sort | Oliver Ploder |
| collection | DOAJ |
| description | Abstract Many researchers and practitioners make heavy use of the least mean squares (LMS) algorithm as an efficient adaptive filter suitable for a multitude of problems. Despite being versatile and efficient, a drawback of this algorithm is that the adaptation rate, i.e. step‐size, has to be chosen very carefully in order to get the desired result (optimum compromise between fast adaptation and low steady state error). This choice was simplified by the invention of the normalised LMS, which bounds the step‐size and guarantees convergence. However, the optimum choice of the normalisation becomes non‐trivial if the system to be approximated is part of a bigger, non‐trivial model, e.g. cascaded filters or linear paths followed by nonlinearities. Such cases usually require approximations or worst‐case estimates in order to yield a normalised update algorithm, which might result in sub‐optimal performance. To counteract this problem, a new class of LMS algorithms which automatically choose their own normalisation terms, the so‐called self normalising LMS, is introduced. The simulations show that this new algorithm not only outperforms state‐of‐the‐art solutions in terms of steady state performance in a cascaded filter scenario but also converges just as fast as all other considered algorithms. |
| first_indexed | 2024-04-10T05:48:22Z |
| format | Article |
| id | doaj.art-2d414fa410474fba8e19d094b02351d9 |
| institution | Directory Open Access Journal |
| issn | 0013-5194 1350-911X |
| language | English |
| last_indexed | 2024-04-10T05:48:22Z |
| publishDate | 2022-06-01 |
| publisher | Wiley |
| record_format | Article |
| series | Electronics Letters |
| spelling | doaj.art-2d414fa410474fba8e19d094b02351d92023-03-05T10:20:33ZengWileyElectronics Letters0013-51941350-911X2022-06-01581249249410.1049/ell2.12498A new class of self‐normalising LMS algorithmsOliver Ploder0Oliver Lang1Thomas Paireder2Christian Motz3Mario Huemer4Christian Doppler Laboratory for Digitally Assisted RF Transceivers for Future Mobile Communications Institute of Signal Processing Johannes Kepler University Linz AustriaInstitute of Signal Processing Johannes Kepler University Linz AustriaChristian Doppler Laboratory for Digitally Assisted RF Transceivers for Future Mobile Communications Institute of Signal Processing Johannes Kepler University Linz AustriaChristian Doppler Laboratory for Digitally Assisted RF Transceivers for Future Mobile Communications Institute of Signal Processing Johannes Kepler University Linz AustriaInstitute of Signal Processing Johannes Kepler University Linz AustriaAbstract Many researchers and practitioners make heavy use of the least mean squares (LMS) algorithm as an efficient adaptive filter suitable for a multitude of problems. Despite being versatile and efficient, a drawback of this algorithm is that the adaptation rate, i.e. step‐size, has to be chosen very carefully in order to get the desired result (optimum compromise between fast adaptation and low steady state error). This choice was simplified by the invention of the normalised LMS, which bounds the step‐size and guarantees convergence. However, the optimum choice of the normalisation becomes non‐trivial if the system to be approximated is part of a bigger, non‐trivial model, e.g. cascaded filters or linear paths followed by nonlinearities. Such cases usually require approximations or worst‐case estimates in order to yield a normalised update algorithm, which might result in sub‐optimal performance. To counteract this problem, a new class of LMS algorithms which automatically choose their own normalisation terms, the so‐called self normalising LMS, is introduced. The simulations show that this new algorithm not only outperforms state‐of‐the‐art solutions in terms of steady state performance in a cascaded filter scenario but also converges just as fast as all other considered algorithms.https://doi.org/10.1049/ell2.12498 |
| spellingShingle | Oliver Ploder Oliver Lang Thomas Paireder Christian Motz Mario Huemer A new class of self‐normalising LMS algorithms Electronics Letters |
| title | A new class of self‐normalising LMS algorithms |
| title_full | A new class of self‐normalising LMS algorithms |
| title_fullStr | A new class of self‐normalising LMS algorithms |
| title_full_unstemmed | A new class of self‐normalising LMS algorithms |
| title_short | A new class of self‐normalising LMS algorithms |
| title_sort | new class of self normalising lms algorithms |
| url | https://doi.org/10.1049/ell2.12498 |
| work_keys_str_mv | AT oliverploder anewclassofselfnormalisinglmsalgorithms AT oliverlang anewclassofselfnormalisinglmsalgorithms AT thomaspaireder anewclassofselfnormalisinglmsalgorithms AT christianmotz anewclassofselfnormalisinglmsalgorithms AT mariohuemer anewclassofselfnormalisinglmsalgorithms AT oliverploder newclassofselfnormalisinglmsalgorithms AT oliverlang newclassofselfnormalisinglmsalgorithms AT thomaspaireder newclassofselfnormalisinglmsalgorithms AT christianmotz newclassofselfnormalisinglmsalgorithms AT mariohuemer newclassofselfnormalisinglmsalgorithms |