Hofmann C, Kellermann W (2018)
Publication Language: English
Publication Type: Book chapter / Article in edited volumes
Publication year: 2018
Publisher: Butterworth-Heinemann
Edited Volumes: Adaptive learning methods for nonlinear system modeling
Pages Range: 71-102
ISBN: 978-0-12-812976-0
URI: http://www.sciencedirect.com/science/article/pii/B9780128129760000051
Linear-In-the-Parameters (LIP) nonlinear filters are categorized as Cascade Models (CMs) (generalizing Hammerstein models), Cascade Group Models (CGMs) (generalizing Hammerstein Group models (HGMs) and including, e.g., Volterra filters) and bilinear cascade models, where the filter output is a bilinear function of the model parameters. Time-domain and partitioned-block frequency-domain adaptation of CGMs and CMs is described and the methods for adapting bilinear cascade models are summarized as variants of the filtered-X adaptation. These models and algorithms are employed to review the Significance-Aware (SA) filtering concept, decomposing the model for the unknown system and the adaptation mechanism into synergetic subsystems to achieve high computational efficiency. In particular, the Serial SA (SSA) and Parallel SA (PSA) decomposition lead to SSA-CMs, PSA-CGMs and a novel PSA filtered-X algorithm. The main concepts described in this chapter are exemplarily compared for the challenging application of nonlinear acoustic echo cancellation. Furthermore, model structure estimation for LIP nonlinear filters based on convex filter combinations is briefly outlined and compared with SA filtering.
APA:
Hofmann, C., & Kellermann, W. (2018). Chapter 4 - Recent advances on LIP nonlinear filters and their applications: Efficient solutions and significance-aware filtering. In Adaptive learning methods for nonlinear system modeling. (pp. 71-102). Butterworth-Heinemann.
MLA:
Hofmann, Christian, and Walter Kellermann. "Chapter 4 - Recent advances on LIP nonlinear filters and their applications: Efficient solutions and significance-aware filtering." Adaptive learning methods for nonlinear system modeling. Butterworth-Heinemann, 2018. 71-102.
BibTeX: Download