The Simplified Wiener LMS Algorithm

Lars Lindbom , Mikael Sternad and Anders Ahlén

6th International Conference on Advances in Communication and Control
(Comcon 6),
Corfu, Greece, June 23-27, 1997.

Outline:

Adaptation algorithms are used for adjusting the coefficients of models, filters and controllers. When fast parameter changes are required, present-day algorithms such as LMS and RLS are severely challenged. The paper presents a new type of tracking algorithm which can, for suitable problems, provide high performance while maintaining LMS complexity.

Abstract:

A new type of tracking algorithm with time-invariant gain is presented. It can be applied for obtaining prediction, filtering or fixed-lag smoothing estimates of time-varying parameters in linear regression models.

The algorithm design constitutes a systematic way of introducing a priori information into LMS-like adaptation laws, using the concept of stochastic hypermodelling of the unknown time-varying parameters. The design equations, which provide the structure and adjustment of the tracking algorithms, are derived from a Wiener filtering perspective.

The simplest variant of the novel class of algorithms, denoted Simplified Wiener LMS (SWLMS), is presented here. The SWLMS algorithm is particularly well suited for tracking of parameters of mobile radio channels. The utility of the algorithm will be demonstrated on a mobile radio channel, where channel coefficients are subject to Rayleigh fading. The tracking scenario refers to the D-AMPS 1900MHz standard.

Related publications:

PhD Thesis by L. Lindbom, presenting the whole class of algorithms.
Master Thesis on the use of a fixed grid of algorithms in D-AMPS 1900.
Conference paper in IEEE ICASSP'93 on an early version of algorithm.
Deterministic modelling of time-variations in digital mobile radio channels.

Source:

Pdf, 600K