Abstract | ||
---|---|---|
In the context of nonlinear systems identification new Affine Projection Algorithm (APA) and NLMS adaptive filters (AFs) are developed over the Simple Multilinear model (SML). Such a model is comprised of a product of linear filters and allows for an exponential decrease in complexity when compared to the complete Volterra model. The MSE surface is developed in terms of data statistical moments and its gradient vector is presented, computing the corresponding Hessian matrix in the sequel. The AFs are generated via stochastic approximations for the data moments and a series of non-trivial derivations resulting in an APA implementation structurally similar to the standard APA recursion. The NLMS algorithm is derived as a particular case. Simulations show good convergence properties when identifying unknown SML and Volterra plants. |
Year | Venue | Field |
---|---|---|
2016 | European Signal Processing Conference | Convergence (routing),Applied mathematics,Mathematical optimization,Nonlinear system,Linear filter,Matrix (mathematics),Hessian matrix,Adaptive filter,Multilinear map,Mathematics,Method of moments (statistics) |
DocType | ISSN | Citations |
Conference | 2076-1465 | 0 |
PageRank | References | Authors |
0.34 | 0 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Felipe C. Pinheiro | 1 | 1 | 1.03 |
Cássio Guimarães Lopes | 2 | 394 | 32.32 |